Original Research

Early adolescent childbearing is associated with a wide range of adverse consequences and restricted life opportunities for young girls and the children they bear.^1-5 Helping adolescents delay early childbearing has long been a goal of healthcare providers, researchers, and policymakers.^6,7 Although the adolescent pregnancy rate in the United States is decreasing in most groups,⁸ it is still disturbingly high, particularly among Hispanic girls.

Most efforts to prevent or delay adolescent pregnancy have been directed at providing birth control, but this intervention is likely to fail if teens are not interested in preventing pregnancy. Although several studies have examined the factors and motivations underlying adolescent contraceptive behavior,^9-13 teen attitudes toward pregnancy are still poorly understood.¹⁴ Adolescents may not share the same negative view of their childbearing as do adults concerned with preventing it. Retrospective studies suggest that as many as 60% to 80% of teenaged pregnancies are “unintended.”^15,16 Other studies examining pregnant and parenting adolescents’ attitudes toward childbearing suggest that the percentage of pregnancies that are truly unintended may be lower than commonly believed.^10,17-19 A significant percentage of never-pregnant adolescents harbor either highly ambivalent or positive attitudes toward early childbearing.^19-21 A better understanding of the factors associated with a desire for pregnancy among adolescents may help health care providers better predict the most at-risk adolescents.

Our study investigated factors associated with adolescent desire for pregnancy among girls seeking reproductive health services at an adolescent clinic. Earlier studies have examined the attitudes of girls presenting for pregnancy testing, or who were already enrolled in prenatal care.⁹ No other studies have specifically looked at the factors associated with nonpregnant adolescents’ attitudes toward pregnancy, or the role of the boyfriend in influencing those attitudes.

Methods

Setting

This study was conducted in an adolescent health clinic within a migrant/community health center in a town with a population of 10,000, 25 miles from a major midwestern city. The clinic was staffed by a nurse practitioner and an adolescent health educator. The full spectrum of adolescent problems are dealt with in this clinic, but most visits are for pregnancy testing, birth control counseling, and checks for sexually transmitted diseases (STDs). Ninety-eight percent of visits are by females. The clinic serves patients aged 12 through 20 years, and has been operating in its current location for 20 years. The clinic offers completely confidential services to those who request it, with no parental notification or consent required.

Subjects

Girls aged 13 to18 years presenting consecutively for reproductive health services were eligible for this study. Girls were excluded if they had delivered a baby in the previous 12 months, had a miscarriage or a therapeutic abortion in the previous 6 months, or were currently using a hormonal method of contraception. A total of 202 girls were initially eligible and all agreed to participate. Because studies have shown that a woman’s attitude toward pregnancy changes once she is aware of being pregnant, we subsequently decided to also exclude those girls already pregnant at the initial visit (n=54), leaving 148 subjects in the final data analysis. All subjects spoke either Spanish or English.

Data Collection

After obtaining informed consent, each girl underwent an extensive, semistructured interview exploring her attitude toward pregnancy, childbearing, and contraceptive use. All interviews were done by 1 of the 2 clinicians working in the adolescent clinic (a 35-year-old adolescent health educator and a 40-year-old obstetric nurse practitioner). Both clinicians were white women who had been working in this clinic for more than 15 years. The health educator was fluent in Spanish. A girl’s attitude toward pregnancy was determined by a series of questions. Other information elicited included ethnicity, age, school attendance, employment status, social habits (alcohol and tobacco use, current dating, current sexual activity), family structure, whether the subject had ever lived away from home for more than 2 weeks, age of her boyfriend, boyfriend’s attitude about pregnancy, and confidentiality of the initial visit. Obstetric and gynecologic histories of each girl were also obtained. After the interview, pregnancy testing was done, if indicated.

Data Analysis

Based on the interview, each girl was categorized as desiring pregnancy, wishing to avoid pregnancy, or being ambivalent about pregnancy. Girls desiring pregnancy were so similar to girls who felt ambivalent about pregnancy that these girls were grouped together in the final analysis and compared against those girls wishing to avoid pregnancy.

Data were analyzed using the SAS statistical program (version 8.0). We used chi-squared testing for unadjusted analysis of factors associated with adolescent attitudes toward pregnancy. Unadjusted associations with a P value greater than .2 were included in multiple logistic regression analysis to adjust for multiple variables, and to calculate odds ratios and 95% confidence intervals. The final multiple logistic regression model included only those variables found to retain significance at P less than .05.

Results

Demographic characteristics of the participants are shown in Table 1. Of the 148 girls, almost all (92%) were currently dating, and most (88%) were sexually active with their partner. Ninety-six percent had never been pregnant previously, 86% had never used hormonal contraception, and 84% had never had a Papanicolaou test or an STD screening.

The mean age of subjects’ boyfriends was 18.4 years, with an age range of 13 to 30 years. One third of the girls lived with both of their biological parents, and 78% lived with at least one biological parent. Three were already married at the time of initial visit. Almost half (46%) described their enrollment visit as confidential.

One hundred seven (56.4%) of the girls were categorized as wishing to avoid pregnancy, 16 girls (19.8%) as desiring pregnancy, and 25 girls (23.7%) as ambivalent about whether they wanted to be pregnant. Unadjusted analysis comparing girls desiring pregnancy with those feeling ambivalent revealed only 1 significant difference: girls desiring pregnancy were more likely to report that their boyfriends wanted a baby. So, in the final analysis, these girls were grouped together, and compared with the girls wishing to avoid pregnancy (Table 1w.

Unadjusted analysis of the 148 subjects is shown in Table 2. Girls wishing to avoid pregnancy differed from girls desiring or ambivalent about pregnancy in 6 different parameters: ethnicity, school attendance, employment status, family structure, time spent away from home, and desire of boyfriend to have a baby. There was no significant association between a girl’s age and her attitude toward pregnancy (P=.48). Notably, the mean age difference between girls and their boyfriends was not significantly associated with desire for pregnancy.

In multivariate analysis of characteristics of the girls themselves, factors significantly associated with a positive attitude toward pregnancy were Hispanic ethnicity, having lived away from home for more than 2 weeks, and having left school Table 3.

When boyfriend characteristics and attitudes were added to the analysis, all subject characteristics ceased to be significant, leaving the perception of the boyfriends’ desire for pregnancy as the only significant variable.

Discussion

We found the strongest predictor of an adolescent girl’s attitude toward pregnancy was her stated belief about whether her boyfriend wanted a baby. In light of the powerful influence of the girl’s perception of her boyfriend’s attitude toward pregnancy, no other factors are significantly associated with her own attitude toward pregnancy. This finding suggests that family physicians and other health care providers working with teenaged girls should include the boyfriend in any discussions aimed at delaying pregnancy.

Girls ambivalent about pregnancy are markedly similar to those desiring it, differing only in the degree to which they believe their boyfriends want a baby. It may be that some of the ambivalence about pregnancy arises from a difference of opinion between the girl and her boyfriend. Girls ambivalent about pregnancy were least likely to know their boyfriend’s opinion on the subject. It may be that young girls who are ambivalent about pregnancy are also those with more limited interpersonal communication skills, making it difficult for them to discuss critical reproductive health issues with their partners. Health care providers may have a role in facilitating improved communication between girls and their partners by specifically addressing partner communication when seeing girls individually, as well as by inviting their partners to be present and more actively involved in clinic visits. Interventions focused solely on providing information about and access to contraception are unlikely to be sufficient in strengthening a girl’s motivation to delay pregnancy. More appropriate and effective interventions may be those that explore the extent to which her partner’s attitudes shape her own critical reproductive health decisions, and encourage greater dialogue between a girl and her partner with respect to contraceptive and childbearing decisions.

Although several studies have been done on the contraceptive behavior of adolescent girls, to our knowledge no other studies have focused on evaluating the influence of the boyfriend’s perceived attitude toward childbearing on nonpregnant adolescent girls’ desire for a child. Our study did not support other studies’ findings showing that girls desiring pregnancy are more likely to have older boyfriends.²² In our study population, neither boyfriend age nor the age difference between the girl and her boyfriend were significantly associated with the girl’s desire to become pregnant.

Limitations

Our study has several limitations. We did not talk with the boyfriends themselves, but instead were limited to what the girls reported about their boyfriends. The girls’ perceptions of their boyfriends’ attitudes toward pregnancy may be more a reflection of the girls’ own desires. Also, we do not know what the girls really thought about pregnancy, only what they reported to us. It may be that more girls desired pregnancy, but were not willing to admit it. Our study did not use a previously validated questionnaire to determine “intendedness” of pregnancy. Because of the difficulty in ascribing motivations to adolescent behavior and reported attitudes, the entire concept of intendedness of pregnancy may not be relevant when discussing adolescent pregnancies.^18,23 However, the semistructured interview used in our study elicited a rich and detailed explanation of attitudes toward such topics as birth control, pregnancy, and influences of family and boyfriend.

Our study results may not be generalizable to other adolescent populations. Our study sample was from a mostly rural area, and the only minority group represented was Hispanic. The vast majority of girls visiting our community health center are from low socioeconomic groups. It is not clear that our results would be true for other ethnicities or for girls from higher socioeconomic levels.

Conclusions

A boyfriend’s desire for a baby is best predictor of an adolescent girl’s attitude toward pregnancy. The most effective interventions may be those that explore the extent to which a boyfriend’s attitude shapes a girl’s critical reproductive health decisions. Primary care providers should include boyfriends in any efforts to delay pregnancy in at-risk adolescent girls and should encourage greater dialogue between the girl and her partner with respect to contraceptive and childbearing decisions.

Acknowledgments

The authors would like to acknowledge the following people for their assistance with this study: Kathy Beamis, for help with data collection; Sherry Holcomb, MS, and Debbi Main, PhD, for help with data analysis; the University of Colorado Primary Care Faculty Development Fellowship group, for review and suggestions.

Related resources

• National Campaign to Prevent Teen Pregnancy http://www.teenpregnancy.org
• National Organization on Adolescent Pregnancy, Parenting, and Prevention http://www.noappp.org
• The Alan Guttmacher Institute Nonprofit organization focused on reproductive health research, policy analysis and public education. http://www.agi-usa.org
• Advocates for Youth Programs designed to help young people make informed decisions about reproductive and sexual health. http://advocatesforyouth.org

Early adolescent childbearing is associated with a wide range of adverse consequences and restricted life opportunities for young girls and the children they bear.^1-5 Helping adolescents delay early childbearing has long been a goal of healthcare providers, researchers, and policymakers.^6,7 Although the adolescent pregnancy rate in the United States is decreasing in most groups,⁸ it is still disturbingly high, particularly among Hispanic girls.

Most efforts to prevent or delay adolescent pregnancy have been directed at providing birth control, but this intervention is likely to fail if teens are not interested in preventing pregnancy. Although several studies have examined the factors and motivations underlying adolescent contraceptive behavior,^9-13 teen attitudes toward pregnancy are still poorly understood.¹⁴ Adolescents may not share the same negative view of their childbearing as do adults concerned with preventing it. Retrospective studies suggest that as many as 60% to 80% of teenaged pregnancies are “unintended.”^15,16 Other studies examining pregnant and parenting adolescents’ attitudes toward childbearing suggest that the percentage of pregnancies that are truly unintended may be lower than commonly believed.^10,17-19 A significant percentage of never-pregnant adolescents harbor either highly ambivalent or positive attitudes toward early childbearing.^19-21 A better understanding of the factors associated with a desire for pregnancy among adolescents may help health care providers better predict the most at-risk adolescents.

Our study investigated factors associated with adolescent desire for pregnancy among girls seeking reproductive health services at an adolescent clinic. Earlier studies have examined the attitudes of girls presenting for pregnancy testing, or who were already enrolled in prenatal care.⁹ No other studies have specifically looked at the factors associated with nonpregnant adolescents’ attitudes toward pregnancy, or the role of the boyfriend in influencing those attitudes.

Methods

Setting

This study was conducted in an adolescent health clinic within a migrant/community health center in a town with a population of 10,000, 25 miles from a major midwestern city. The clinic was staffed by a nurse practitioner and an adolescent health educator. The full spectrum of adolescent problems are dealt with in this clinic, but most visits are for pregnancy testing, birth control counseling, and checks for sexually transmitted diseases (STDs). Ninety-eight percent of visits are by females. The clinic serves patients aged 12 through 20 years, and has been operating in its current location for 20 years. The clinic offers completely confidential services to those who request it, with no parental notification or consent required.

Subjects

Girls aged 13 to18 years presenting consecutively for reproductive health services were eligible for this study. Girls were excluded if they had delivered a baby in the previous 12 months, had a miscarriage or a therapeutic abortion in the previous 6 months, or were currently using a hormonal method of contraception. A total of 202 girls were initially eligible and all agreed to participate. Because studies have shown that a woman’s attitude toward pregnancy changes once she is aware of being pregnant, we subsequently decided to also exclude those girls already pregnant at the initial visit (n=54), leaving 148 subjects in the final data analysis. All subjects spoke either Spanish or English.

Data Collection

After obtaining informed consent, each girl underwent an extensive, semistructured interview exploring her attitude toward pregnancy, childbearing, and contraceptive use. All interviews were done by 1 of the 2 clinicians working in the adolescent clinic (a 35-year-old adolescent health educator and a 40-year-old obstetric nurse practitioner). Both clinicians were white women who had been working in this clinic for more than 15 years. The health educator was fluent in Spanish. A girl’s attitude toward pregnancy was determined by a series of questions. Other information elicited included ethnicity, age, school attendance, employment status, social habits (alcohol and tobacco use, current dating, current sexual activity), family structure, whether the subject had ever lived away from home for more than 2 weeks, age of her boyfriend, boyfriend’s attitude about pregnancy, and confidentiality of the initial visit. Obstetric and gynecologic histories of each girl were also obtained. After the interview, pregnancy testing was done, if indicated.

Data Analysis

Based on the interview, each girl was categorized as desiring pregnancy, wishing to avoid pregnancy, or being ambivalent about pregnancy. Girls desiring pregnancy were so similar to girls who felt ambivalent about pregnancy that these girls were grouped together in the final analysis and compared against those girls wishing to avoid pregnancy.

Data were analyzed using the SAS statistical program (version 8.0). We used chi-squared testing for unadjusted analysis of factors associated with adolescent attitudes toward pregnancy. Unadjusted associations with a P value greater than .2 were included in multiple logistic regression analysis to adjust for multiple variables, and to calculate odds ratios and 95% confidence intervals. The final multiple logistic regression model included only those variables found to retain significance at P less than .05.

Results

Demographic characteristics of the participants are shown in Table 1. Of the 148 girls, almost all (92%) were currently dating, and most (88%) were sexually active with their partner. Ninety-six percent had never been pregnant previously, 86% had never used hormonal contraception, and 84% had never had a Papanicolaou test or an STD screening.

The mean age of subjects’ boyfriends was 18.4 years, with an age range of 13 to 30 years. One third of the girls lived with both of their biological parents, and 78% lived with at least one biological parent. Three were already married at the time of initial visit. Almost half (46%) described their enrollment visit as confidential.

One hundred seven (56.4%) of the girls were categorized as wishing to avoid pregnancy, 16 girls (19.8%) as desiring pregnancy, and 25 girls (23.7%) as ambivalent about whether they wanted to be pregnant. Unadjusted analysis comparing girls desiring pregnancy with those feeling ambivalent revealed only 1 significant difference: girls desiring pregnancy were more likely to report that their boyfriends wanted a baby. So, in the final analysis, these girls were grouped together, and compared with the girls wishing to avoid pregnancy (Table 1w.

Unadjusted analysis of the 148 subjects is shown in Table 2. Girls wishing to avoid pregnancy differed from girls desiring or ambivalent about pregnancy in 6 different parameters: ethnicity, school attendance, employment status, family structure, time spent away from home, and desire of boyfriend to have a baby. There was no significant association between a girl’s age and her attitude toward pregnancy (P=.48). Notably, the mean age difference between girls and their boyfriends was not significantly associated with desire for pregnancy.

In multivariate analysis of characteristics of the girls themselves, factors significantly associated with a positive attitude toward pregnancy were Hispanic ethnicity, having lived away from home for more than 2 weeks, and having left school Table 3.

When boyfriend characteristics and attitudes were added to the analysis, all subject characteristics ceased to be significant, leaving the perception of the boyfriends’ desire for pregnancy as the only significant variable.

Discussion

We found the strongest predictor of an adolescent girl’s attitude toward pregnancy was her stated belief about whether her boyfriend wanted a baby. In light of the powerful influence of the girl’s perception of her boyfriend’s attitude toward pregnancy, no other factors are significantly associated with her own attitude toward pregnancy. This finding suggests that family physicians and other health care providers working with teenaged girls should include the boyfriend in any discussions aimed at delaying pregnancy.

Girls ambivalent about pregnancy are markedly similar to those desiring it, differing only in the degree to which they believe their boyfriends want a baby. It may be that some of the ambivalence about pregnancy arises from a difference of opinion between the girl and her boyfriend. Girls ambivalent about pregnancy were least likely to know their boyfriend’s opinion on the subject. It may be that young girls who are ambivalent about pregnancy are also those with more limited interpersonal communication skills, making it difficult for them to discuss critical reproductive health issues with their partners. Health care providers may have a role in facilitating improved communication between girls and their partners by specifically addressing partner communication when seeing girls individually, as well as by inviting their partners to be present and more actively involved in clinic visits. Interventions focused solely on providing information about and access to contraception are unlikely to be sufficient in strengthening a girl’s motivation to delay pregnancy. More appropriate and effective interventions may be those that explore the extent to which her partner’s attitudes shape her own critical reproductive health decisions, and encourage greater dialogue between a girl and her partner with respect to contraceptive and childbearing decisions.

Although several studies have been done on the contraceptive behavior of adolescent girls, to our knowledge no other studies have focused on evaluating the influence of the boyfriend’s perceived attitude toward childbearing on nonpregnant adolescent girls’ desire for a child. Our study did not support other studies’ findings showing that girls desiring pregnancy are more likely to have older boyfriends.²² In our study population, neither boyfriend age nor the age difference between the girl and her boyfriend were significantly associated with the girl’s desire to become pregnant.

Limitations

Our study has several limitations. We did not talk with the boyfriends themselves, but instead were limited to what the girls reported about their boyfriends. The girls’ perceptions of their boyfriends’ attitudes toward pregnancy may be more a reflection of the girls’ own desires. Also, we do not know what the girls really thought about pregnancy, only what they reported to us. It may be that more girls desired pregnancy, but were not willing to admit it. Our study did not use a previously validated questionnaire to determine “intendedness” of pregnancy. Because of the difficulty in ascribing motivations to adolescent behavior and reported attitudes, the entire concept of intendedness of pregnancy may not be relevant when discussing adolescent pregnancies.^18,23 However, the semistructured interview used in our study elicited a rich and detailed explanation of attitudes toward such topics as birth control, pregnancy, and influences of family and boyfriend.

Our study results may not be generalizable to other adolescent populations. Our study sample was from a mostly rural area, and the only minority group represented was Hispanic. The vast majority of girls visiting our community health center are from low socioeconomic groups. It is not clear that our results would be true for other ethnicities or for girls from higher socioeconomic levels.

Conclusions

A boyfriend’s desire for a baby is best predictor of an adolescent girl’s attitude toward pregnancy. The most effective interventions may be those that explore the extent to which a boyfriend’s attitude shapes a girl’s critical reproductive health decisions. Primary care providers should include boyfriends in any efforts to delay pregnancy in at-risk adolescent girls and should encourage greater dialogue between the girl and her partner with respect to contraceptive and childbearing decisions.

Acknowledgments

The authors would like to acknowledge the following people for their assistance with this study: Kathy Beamis, for help with data collection; Sherry Holcomb, MS, and Debbi Main, PhD, for help with data analysis; the University of Colorado Primary Care Faculty Development Fellowship group, for review and suggestions.

Related resources

• National Campaign to Prevent Teen Pregnancy http://www.teenpregnancy.org
• National Organization on Adolescent Pregnancy, Parenting, and Prevention http://www.noappp.org
• The Alan Guttmacher Institute Nonprofit organization focused on reproductive health research, policy analysis and public education. http://www.agi-usa.org
• Advocates for Youth Programs designed to help young people make informed decisions about reproductive and sexual health. http://advocatesforyouth.org

Early adolescent childbearing is associated with a wide range of adverse consequences and restricted life opportunities for young girls and the children they bear.^1-5 Helping adolescents delay early childbearing has long been a goal of healthcare providers, researchers, and policymakers.^6,7 Although the adolescent pregnancy rate in the United States is decreasing in most groups,⁸ it is still disturbingly high, particularly among Hispanic girls.

Most efforts to prevent or delay adolescent pregnancy have been directed at providing birth control, but this intervention is likely to fail if teens are not interested in preventing pregnancy. Although several studies have examined the factors and motivations underlying adolescent contraceptive behavior,^9-13 teen attitudes toward pregnancy are still poorly understood.¹⁴ Adolescents may not share the same negative view of their childbearing as do adults concerned with preventing it. Retrospective studies suggest that as many as 60% to 80% of teenaged pregnancies are “unintended.”^15,16 Other studies examining pregnant and parenting adolescents’ attitudes toward childbearing suggest that the percentage of pregnancies that are truly unintended may be lower than commonly believed.^10,17-19 A significant percentage of never-pregnant adolescents harbor either highly ambivalent or positive attitudes toward early childbearing.^19-21 A better understanding of the factors associated with a desire for pregnancy among adolescents may help health care providers better predict the most at-risk adolescents.

Our study investigated factors associated with adolescent desire for pregnancy among girls seeking reproductive health services at an adolescent clinic. Earlier studies have examined the attitudes of girls presenting for pregnancy testing, or who were already enrolled in prenatal care.⁹ No other studies have specifically looked at the factors associated with nonpregnant adolescents’ attitudes toward pregnancy, or the role of the boyfriend in influencing those attitudes.

Methods

Setting

This study was conducted in an adolescent health clinic within a migrant/community health center in a town with a population of 10,000, 25 miles from a major midwestern city. The clinic was staffed by a nurse practitioner and an adolescent health educator. The full spectrum of adolescent problems are dealt with in this clinic, but most visits are for pregnancy testing, birth control counseling, and checks for sexually transmitted diseases (STDs). Ninety-eight percent of visits are by females. The clinic serves patients aged 12 through 20 years, and has been operating in its current location for 20 years. The clinic offers completely confidential services to those who request it, with no parental notification or consent required.

Subjects

Girls aged 13 to18 years presenting consecutively for reproductive health services were eligible for this study. Girls were excluded if they had delivered a baby in the previous 12 months, had a miscarriage or a therapeutic abortion in the previous 6 months, or were currently using a hormonal method of contraception. A total of 202 girls were initially eligible and all agreed to participate. Because studies have shown that a woman’s attitude toward pregnancy changes once she is aware of being pregnant, we subsequently decided to also exclude those girls already pregnant at the initial visit (n=54), leaving 148 subjects in the final data analysis. All subjects spoke either Spanish or English.

Data Collection

After obtaining informed consent, each girl underwent an extensive, semistructured interview exploring her attitude toward pregnancy, childbearing, and contraceptive use. All interviews were done by 1 of the 2 clinicians working in the adolescent clinic (a 35-year-old adolescent health educator and a 40-year-old obstetric nurse practitioner). Both clinicians were white women who had been working in this clinic for more than 15 years. The health educator was fluent in Spanish. A girl’s attitude toward pregnancy was determined by a series of questions. Other information elicited included ethnicity, age, school attendance, employment status, social habits (alcohol and tobacco use, current dating, current sexual activity), family structure, whether the subject had ever lived away from home for more than 2 weeks, age of her boyfriend, boyfriend’s attitude about pregnancy, and confidentiality of the initial visit. Obstetric and gynecologic histories of each girl were also obtained. After the interview, pregnancy testing was done, if indicated.

Data Analysis

Based on the interview, each girl was categorized as desiring pregnancy, wishing to avoid pregnancy, or being ambivalent about pregnancy. Girls desiring pregnancy were so similar to girls who felt ambivalent about pregnancy that these girls were grouped together in the final analysis and compared against those girls wishing to avoid pregnancy.

Data were analyzed using the SAS statistical program (version 8.0). We used chi-squared testing for unadjusted analysis of factors associated with adolescent attitudes toward pregnancy. Unadjusted associations with a P value greater than .2 were included in multiple logistic regression analysis to adjust for multiple variables, and to calculate odds ratios and 95% confidence intervals. The final multiple logistic regression model included only those variables found to retain significance at P less than .05.

Results

Demographic characteristics of the participants are shown in Table 1. Of the 148 girls, almost all (92%) were currently dating, and most (88%) were sexually active with their partner. Ninety-six percent had never been pregnant previously, 86% had never used hormonal contraception, and 84% had never had a Papanicolaou test or an STD screening.

The mean age of subjects’ boyfriends was 18.4 years, with an age range of 13 to 30 years. One third of the girls lived with both of their biological parents, and 78% lived with at least one biological parent. Three were already married at the time of initial visit. Almost half (46%) described their enrollment visit as confidential.

One hundred seven (56.4%) of the girls were categorized as wishing to avoid pregnancy, 16 girls (19.8%) as desiring pregnancy, and 25 girls (23.7%) as ambivalent about whether they wanted to be pregnant. Unadjusted analysis comparing girls desiring pregnancy with those feeling ambivalent revealed only 1 significant difference: girls desiring pregnancy were more likely to report that their boyfriends wanted a baby. So, in the final analysis, these girls were grouped together, and compared with the girls wishing to avoid pregnancy (Table 1w.

Unadjusted analysis of the 148 subjects is shown in Table 2. Girls wishing to avoid pregnancy differed from girls desiring or ambivalent about pregnancy in 6 different parameters: ethnicity, school attendance, employment status, family structure, time spent away from home, and desire of boyfriend to have a baby. There was no significant association between a girl’s age and her attitude toward pregnancy (P=.48). Notably, the mean age difference between girls and their boyfriends was not significantly associated with desire for pregnancy.

In multivariate analysis of characteristics of the girls themselves, factors significantly associated with a positive attitude toward pregnancy were Hispanic ethnicity, having lived away from home for more than 2 weeks, and having left school Table 3.

When boyfriend characteristics and attitudes were added to the analysis, all subject characteristics ceased to be significant, leaving the perception of the boyfriends’ desire for pregnancy as the only significant variable.

Discussion

We found the strongest predictor of an adolescent girl’s attitude toward pregnancy was her stated belief about whether her boyfriend wanted a baby. In light of the powerful influence of the girl’s perception of her boyfriend’s attitude toward pregnancy, no other factors are significantly associated with her own attitude toward pregnancy. This finding suggests that family physicians and other health care providers working with teenaged girls should include the boyfriend in any discussions aimed at delaying pregnancy.

Girls ambivalent about pregnancy are markedly similar to those desiring it, differing only in the degree to which they believe their boyfriends want a baby. It may be that some of the ambivalence about pregnancy arises from a difference of opinion between the girl and her boyfriend. Girls ambivalent about pregnancy were least likely to know their boyfriend’s opinion on the subject. It may be that young girls who are ambivalent about pregnancy are also those with more limited interpersonal communication skills, making it difficult for them to discuss critical reproductive health issues with their partners. Health care providers may have a role in facilitating improved communication between girls and their partners by specifically addressing partner communication when seeing girls individually, as well as by inviting their partners to be present and more actively involved in clinic visits. Interventions focused solely on providing information about and access to contraception are unlikely to be sufficient in strengthening a girl’s motivation to delay pregnancy. More appropriate and effective interventions may be those that explore the extent to which her partner’s attitudes shape her own critical reproductive health decisions, and encourage greater dialogue between a girl and her partner with respect to contraceptive and childbearing decisions.

Although several studies have been done on the contraceptive behavior of adolescent girls, to our knowledge no other studies have focused on evaluating the influence of the boyfriend’s perceived attitude toward childbearing on nonpregnant adolescent girls’ desire for a child. Our study did not support other studies’ findings showing that girls desiring pregnancy are more likely to have older boyfriends.²² In our study population, neither boyfriend age nor the age difference between the girl and her boyfriend were significantly associated with the girl’s desire to become pregnant.

Limitations

Our study has several limitations. We did not talk with the boyfriends themselves, but instead were limited to what the girls reported about their boyfriends. The girls’ perceptions of their boyfriends’ attitudes toward pregnancy may be more a reflection of the girls’ own desires. Also, we do not know what the girls really thought about pregnancy, only what they reported to us. It may be that more girls desired pregnancy, but were not willing to admit it. Our study did not use a previously validated questionnaire to determine “intendedness” of pregnancy. Because of the difficulty in ascribing motivations to adolescent behavior and reported attitudes, the entire concept of intendedness of pregnancy may not be relevant when discussing adolescent pregnancies.^18,23 However, the semistructured interview used in our study elicited a rich and detailed explanation of attitudes toward such topics as birth control, pregnancy, and influences of family and boyfriend.

Our study results may not be generalizable to other adolescent populations. Our study sample was from a mostly rural area, and the only minority group represented was Hispanic. The vast majority of girls visiting our community health center are from low socioeconomic groups. It is not clear that our results would be true for other ethnicities or for girls from higher socioeconomic levels.

Conclusions

A boyfriend’s desire for a baby is best predictor of an adolescent girl’s attitude toward pregnancy. The most effective interventions may be those that explore the extent to which a boyfriend’s attitude shapes a girl’s critical reproductive health decisions. Primary care providers should include boyfriends in any efforts to delay pregnancy in at-risk adolescent girls and should encourage greater dialogue between the girl and her partner with respect to contraceptive and childbearing decisions.

Acknowledgments

The authors would like to acknowledge the following people for their assistance with this study: Kathy Beamis, for help with data collection; Sherry Holcomb, MS, and Debbi Main, PhD, for help with data analysis; the University of Colorado Primary Care Faculty Development Fellowship group, for review and suggestions.

Related resources

• National Campaign to Prevent Teen Pregnancy http://www.teenpregnancy.org
• National Organization on Adolescent Pregnancy, Parenting, and Prevention http://www.noappp.org
• The Alan Guttmacher Institute Nonprofit organization focused on reproductive health research, policy analysis and public education. http://www.agi-usa.org
• Advocates for Youth Programs designed to help young people make informed decisions about reproductive and sexual health. http://advocatesforyouth.org

Asthma, a chronic inflammatory disorder of the airways, affects approximately 17.3 million people in the United States, with more than 5000 deaths per year attributed to the disorder.¹ National Institutes of Health (NIH) guidelines for asthma management recommend that patients with persistent asthma (ie, patients having symptoms more than twice weekly) receive daily medication for long-term control of asthma.¹ These guidelines consider inhaled corticosteroids the most effective controller medication for persistent asthma.

Recently, leukotriene modifiers have been introduced as a new class of asthma medications. NIH guidelines include leukotriene modifiers as potential alternative controller medications for mild, persistent asthma.¹ However, the guidelines also state that additional data are needed to establish the role and position of leukotriene modifiers in asthma therapy. To date, there have been few published reports of placebo-controlled clinical trials comparing the efficacy, safety, and quality of life in asthma patients treated with an inhaled corticosteroid versus a leukotriene modifier.

This study assessed the clinical benefits of an inhaled corticosteroid and a leukotriene modifier as first-line treatment for persistent asthma in patients who were symptomatic when using short-acting b2-agonists alone.

Methods

Patients

Patients aged 12 years old or older with asthma² were eligible if they had used a short-acting b2-agonist (either scheduled or as-needed) for at least 6 weeks preceding the study, if they had a 1-second forced expiratory volume (FEV₁) between 50% to 80% of predicted values, ^3-5 and if they demonstrated reversibility of FEV₁ (Ž12% increase within 30 minutes after inhaling 180 mg albuterol). Patients were excluded from the study if they had life-threatening asthma, significant and uncontrolled diseases (eg, chronic obstructive pulmonary disease, diabetes, coronary disease), used tobacco products within the preceding year, or had a smoking history of more than 10 pack-years. Patients were also excluded if they received any systemic corticosteroid within 6 months of screening, any inhaled corticosteroid within 1 month of screening, or any leukotriene modifier within 1 week of screening.

Study Design

This randomized double-blind, double-dummy, parallel-group study was conducted at 34 sites in the United States. All patients (or their parent or guardian) gave written informed consent before entering the study. The study was approved by institutional review boards for each site.

Eligible patients entered an 8-to 14-day run-in period to establish baseline respiratory function. Patients who met the study criteria were randomly assigned to receive 1 of the following double-blind, double-dummy treatments every morning and evening for 12 weeks: (1) 88 mg fluticasone propionate (FP [Flovent Inhalation Aerosol], 2 puffs of 44 mg each) administered by metered-dose inhaler (MDI) and a placebo capsule; (2) oral zafirlukast 20 mg (Accolate over-encapsulated tablet) and 2 puffs of placebo by MDI; or (3) a placebo capsule and 2 puffs of placebo by MDI. The zafirlukast dosage (20 mg twice daily) is the currently recommended dosage for patients aged 12 years or older, and 88 mg FP twice daily is the lowest FP dosage delivered by MDI approved for use in adolescents and adults. Treatment assignments were generated by computer in blocks of 6 so that each of the 3 treatments was represented twice in random order.

Before initiating the study, comparative dissolution testing was performed to determine whether over-encapsulation would affect the bioequivalence of the blinded zafirlukast tablets. The dissolution profiles were the same for the over-encapsulated (blinded drug) and the trade product. Empty capsules were also tested, and complete dissolution and disintegration of the capsules occurred within 5 minutes.

Patients were instructed to inhale 2 puffs from their study MDI each morning and evening and to swallow 1 study capsule each morning at least 1 hour before breakfast, and each evening at least 2 hours following dinner. The timing of administration of study capsules was based on the limited absorption of zafirlukast when given with food. Patients were instructed to rinse their mouths with water after using the MDI. Patients were not allowed to take any asthma medication during the study other than the study medication, albuterol (Ventolin Inhalation Aerosol) as needed for symptom relief, or oral or parenteral corticosteroids for asthma exacerbations. The use of medications that interact with inhaled corticosteroids or leukotriene modifiers was not allowed. Patients who experienced 2 exacerbations (eg, required a second burst of oral corticosteroids) or received oral or parenteral corticosteroids for more than 14 consecutive days were withdrawn from the study.

Assessments of Efficacy and Safety

The primary physiologic efficacy measure was morning predose FEV₁. Other efficacy measures included patient-recorded morning (am) and evening (pm) peak expiratory flow (PEF). Patient-oriented outcomes included asthma symptom scores, percentage of symptom-free days, albuterol use, percentage of albuterol-free days, and nighttime awakenings due to asthma symptoms; physicians’ overall assessment of medication effectiveness, and patient questionnaires concerning asthma-specific quality of life, productivity, and satisfaction with study medication. Safety assessments included the incidences of asthma exacerbations, adverse events, and pre-defined, clinically significant, laboratory test results.

FEV₁ was assessed at each clinic visit between 6 am and 10 am.⁶ Morning doses of study medications were withheld before each clinic visit, and the visit was rescheduled if the patient used albuterol within 6 hours preceding the visit. Before taking study medications, patients recorded on a diary card any asthma symptoms experienced, using a 6-point scale (where 0 = no symptoms and 5 = symptoms caused severe discomfort and prevented normal activities). After recording symptoms, patients measured and recorded their PEF, along with the number of puffs of albuterol used and the number of nighttime awakenings due to asthma. Patients reported study medication compliance daily on the diary card. Adverse events, concurrent medications, and diary cards were reviewed at each clinic visit. Laboratory tests were performed at the beginning of the run-in period and at the patient’s last visit.

At the last clinic visit, following a review of diary cards and the status of each patient, physicians rated the overall effectiveness of each patient’s blinded study medication using a 4-point scale (where 0 = ineffective and 3 = very effective). At weeks 2 and 12, patients assessed their overall satisfaction with study medication using a 7-point scale (where 0= very dissatisfied and 6 = very satisfied) and completed the Asthma Quality of Life Questionnaire (AQLQ),⁷ which measures the impact of asthma on patients’ daily functioning. At weeks 2, 4, 6, 8, and 12, patients documented their productivity (days missed from work or school and days with symptoms at work or school) over the previous 2 weeks.

Statistical Analyses

The sample size (Ž100 patients completing each treatment) provided Ž80% power to detect a difference in FEV₁ of 0.25 L between treatment groups, according to a 2-sample t test with a significance level of 0.05 and a standard deviation of 0.55 L.

All patients randomly assigned to the study drug were included in the safety analyses. Nine patients at 1 site were excluded from the efficacy analyses because of significant deviations from good clinical practice standards.

Change from baseline was calculated at each clinic visit and at the last available observation for FEV₁. For each subject-recorded assessment from the diary card, weekly means were calculated for each patient, and change from baseline (baseline = mean of last 7 days’ data from run-in) was calculated at each week and at the endpoint (endpoint = mean of last available weekly data). Onset of effect was evaluated as change in daily am PEF from day of randomization. Changes from baseline in global AQLQ and the individual domain scores were computed at the endpoint (endpoint = last available questionnaire). Analyses of covariance adjusted for investigator effect and included the baseline value as a covariate. Analysis of covariance F tests were used to evaluate treatment differences in all change from baseline values.

Cochran-Mantel-Haenszel tests, controlling for investigator, were used to determine treatment differences in physicians’ overall assessment of efficacy (at endpoint), in patient satisfaction with study medication (at week 2 and endpoint), and in the number of patients in each treatment group with at least 1 asthma exacerbation requiring oral or parenteral corticosteroids. Van Elteren tests were used to assess patient productivity.

Results

Patients

A total of 338 patients, 12 to 75 years of age, were randomly assigned to treatment (113, FP; 111, zafirlukast; 114, placebo). Equal numbers of men and women were enrolled in the study, and the majority of patients were non-Hispanic white (86%) or African American (10%). Baseline pulmonary function (% of predicted FEV₁) was similar across treatment groups (66%-67% predicted). The majority of patients in each treatment group had moderate asthma (77%, FP; 82%, zafirlukast; 18%, placebo), and most patients in each treatment group had asthma diagnosed for at least 10 years (64%-73%). A post hoc survey of patients revealed that the majority of respondents (149 of 286 patients; 52%) were recruited from primary care practices for the purpose of the study or were treating their asthma with over-the-counter bronchodilators. Similar numbers of patients in each group completed the study (81%-86% of patients). The most common reasons for discontinuation were lack of efficacy (2%-5%) and withdrawn consent (3%-5%). Median compliance was 93% in each group for both inhaled and oral study medication.

Efficacy

Treatment with FP significantly improved pulmonary function more than zafirlukast or placebo. At the endpoint Table 1 and at individual points as early as day 4 ( Figure 1, mean improvements in FEV₁ and in am PF and pm PEF were significantly greater in the FP group compared with the zafirlukast or placebo group (P <.05). When stratified by baseline pulmonary function (% predicted FEV₁), mean changes in FEV₁ and pm PEF were significantly greater in FP patients than in zafirlukast or placebo patients, regardless of baseline pulmonary function Table 2.

At the endpoint, treatment with FP significantly improved all clinical parameters (mean symptom scores, percentages of symptom-free and albuterol-free days, albuterol use, and nighttime awakenings due to asthma) compared with placebo (P≤.006, Table 1. In contrast, treatment with zafirlukast resulted in significant improvements in albuterol use and the percentage of symptom-free and albuterol-free days (P≤.040, vs placebo), but not in mean symptom scores or the number of nighttime awakenings. Compared with zafirlukast at the endpoint, FP produced significantly greater improvements in mean symptom scores, the percentages of symptom-free and albuterol-free days, albuterol use, and the number of nighttime awakenings due to asthma (P≤.038). When compared with placebo at individual time points, FP produced significantly greater improvements in each clinical parameter at most or all time points (P≤.039). In contrast, treatment with zafirlukast resulted in fewer time points with significantly different improvements in these parameters compared with placebo. Compared with zafirlukast, FP produced significantly greater improvements in mean symptom scores, the percentages of symptom-free and albuterol-free days, albuterol use, and nighttime awakenings due to asthma at the majority of time points from week 2 through week 12 (P≤.050).

At the endpoint, physicians’ overall assessments of study medication efficacy and patients’ overall satisfaction with study medication both significantly favored FP over zafirlukast (P≤.025) or placebo (P <.001; Figure 2. More physicians rated FP “effective” or “very effective” compared with zafirlukast or placebo. Likewise, more FP patients were “satisfied” or “very satisfied” with their study medication than were patients taking zafirlukast or placebo. The patient productivity questionnaire data showed that the mean number of days that patients attended work or school with asthma symptoms was significantly higher in the zafirlukast and placebo groups than the FP group (3.8, 4.4, and 1.8 days, respectively, P≤.03).

At the endpoint, treatment with FP produced significantly greater improvements in global and individual domain AQLQ scores compared with zafirlukast or placebo (P≤.033; Figure 3. Furthermore, treatment differences between FP and placebo for global and most individual domain AQLQ scores are considered clinically meaningful (Ž0.5).⁸ In contrast, zafirlukast did not achieve clinically meaningful differences compared with placebo for any AQLQ endpoint.

Safety

The percentages of patients who experienced at least 1 adverse event were similar across treatment groups (67% to 72%). More FP patients reported sinusitis (12%, vs 4% with zafirlukast and 4% with placebo), and more zafirlukast patients reported chest congestion (5%, vs <1% with FP and 0 with placebo). Oropharyngeal candidiasis was reported for 2 placebo recipients and 3 FP patients. Few patients experienced adverse events potentially related to study medication (12% to 13% of each group). The most common adverse events considered potentially related to study medication were throat irritation (FP: 4% of patients; zafirlukast and placebo: 3% each) and headache (FP: 3%; placebo and zafirlukast: 2% each). Two patients in the FP group and 1 patient in the zafirlukast group and 1 patient in the placebo group discontinued study treatment because of adverse events possibly related to study medication (FP: diarrhea, coughing and hoarseness; zafirlukast: increase in g-glutamyltransferase [GGT; other hepatic enzyme levels were not elevated]; placebo: lightheadedness and shakiness). The only adverse events requiring hospitalization occurred in the placebo group. Few clinically significant laboratory abnormalities ( 3% for any assay) were reported; none were considered related to study medication. Fewer FP patients experienced asthma exacerbations that required treatment with oral corticosteroids compared with zafirlukast or placebo patients (4%, 12%, and 10% of patients, respectively); treatment differences, however, were not statistically significant.

Discussion

The goals of asthma therapy are to prevent asthma symptoms, maintain normal activity levels, normalize lung function, meet patients’ expectations and satisfaction, and provide optimal therapy with minimal adverse effects.¹ For patients with persistent asthma, daily controller medications are recommended.¹ The results of our study indicate that, compared with zafirlukast or placebo, treatment with FP resulted in greater improvements in patients’ symptoms, quality of life, and overall satisfaction, as well as less albuterol use, fewer nighttime awakenings, and better pulmonary function. Furthermore, fewer FP patients experienced an asthma exacerbation compared with zafirlukast or placebo patients. Improvements in the zafirlukast group were generally similar to improvements reported in previous placebo-controlled trials with zafirlukast.^9,10

In our study, differences between FP and zafirlukast were seen early, and those differences generally persisted throughout the 12-week treatment. Treatment effects that favored FP were seen for all clinical endpoints by week 2, while differences in pulmonary function that favored FP were noted by day 4 of treatment. The rapid improvements noted with FP in our study are consistent with previously published results.¹¹ In the current study, patients treated with FP continued to improve with each subsequent week of treatment, while improvements with zafirlukast generally peaked after 2 weeks of treatment. Although the percentage of symptom-free days in the zafirlukast group appeared to increase in the last 4 weeks of the study, this may be related to the higher withdrawal rate during this period. Patient-rated outcomes support the clinical findings. Significant differences between FP and zafirlukast occurred by week 2 with quality of life, patient satisfaction, and days worked with asthma symptoms. The treatment effects noted in this study may be related to the rapid onset of action of fluticasone¹¹ and may result in improved compliance and better long-term asthma control.

Preventing recurrent exacerbations and minimizing the need for emergency department visits or hospitalizations is an important goal in asthma therapy.¹ The low incidence of exacerbations in the FP group in our study is consistent with data from previous clinical trials.^12-14 With regard to the frequency of asthma exacerbations in patients receiving leukotriene modifiers, available data demonstrate that zileuton, zafirlukast, and montelukast each significantly decrease the risk of asthma exacerbations compared with placebo.^10,15-17 Although leukotriene modifiers are superior to placebo in reducing the risk of asthma exacerbations, the results of our study indicate that patients receiving an inhaled corticosteroid had fewer asthma exacerbations than those receiving a leukotriene receptor antagonist. These data are in agreement with an earlier placebo-controlled, comparative study.¹⁷

A significant placebo effect was seen with FEV₁ in our study and is similar to observations from other FP studies.^18,19 In our study, the placebo effect was larger in patients with a percent predicted FEV¹ of more than 70% at baseline. One possible explanation for this is regression to the mean, since patients typically present for care when their symptoms are worst. Another is that milder patients are affected more by the additional medical attention received in clinical studies or that participation in a clinical study influenced patients to avoid asthma-triggering events. The placebo effect was strongest with the spirometric endpoints, for reasons that are unclear. Regardless of the magnitude of the placebo response, the differences in our study between placebo and each active treatment are consistent with the differences seen in other studies involving FP or zafirlukast.^18,20

Because asthma involves chronic inflammation of the airways, the NIH guidelines recommend long-term controller medications with anti-inflammatory effects. In the present study, an inhaled cortico-steroid and a leukotriene modifier both led to improvements in pulmonary function. However, FP continued to improve pulmonary function throughout the 12-week study, while the benefits of zafirlukast tended to plateau by week 2 of treatment. Although the precise mechanism of action of glucocorticoids in asthma is unknown, the greater degree of effectiveness seen with FP in the present study may be related to the multiple effects of inhaled corticosteroids on airway inflammation.²¹ A post hoc analysis of the patient population stratified by baseline pulmonary function (% predicted FEV₁ of 50% to 70% or 70.1% to 80%) showed that, regardless of baseline lung function, FP was more effective than zafirlukast or placebo with regard to improvement in albuterol use and pulmonary function. These data are consistent with asthma guidelines, which recommend inhaled corticosteroids as first-line treatment for all levels of persistent asthma.¹

Limitations

The design of our study has several potential limitations. First, several studies indicate that dosages of zafirlukast higher than 20 mg twice daily may result in greater pharmacologic results.^22-26 Our study, however, was performed in the United States, where this is the only approved dosage for zafirlukast. Second, outcomes were measured at the end of the dosing interval, before administration of morning dosages of study medication. The timing of these tests may have introduced a potential bias in favor of FP, since there is evidence to support bronchodilatory activity with leukotriene modifiers.²⁷ The lesser effect of zafirlukast could be explained if the drug has a duration of action that is shorter than the recommended 12-hour dosing interval. However, the procedures and timing used in this study are similar to those used in other studies of leukotriene modifiers.^9,10,17 Third, a large number of efficacy measures were collected at multiple time points, which potentially increases the likelihood of finding significant treatment differences. However, many of the differences noted in this study were highly significant, and all of the measures were in agreement. This suggests that treatment differences seen in this study were valid. Finally, zafirlukast tablets were over-encapsulated for the purpose of blinding. The over-encapsulation did not appear to affect the performance of zafirlukast, because over-encapsulated and nonencapsulated tablets exhibited similar dissolution profiles. This conclusion is reinforced by the fact that the efficacy profile of zafirlukast in our study is similar to that seen in previous studies.^9,10

Conclusions

Treatment with FP 88 mg twice daily was more effective than zafirlukast 20 mg twice daily or placebo in improving asthma control and patient-rated outcomes in patients with persistent asthma who were previously receiving short-acting b2-agonists alone.

Acknowledgments

We acknowledge the important contributions made by the clinical investigators in this study: Thomas Bell, MD, Missoula, MT; Robert Berkowitz, MD, Atlanta, GA; David Bernstein, MD, Cincinnati, OH; Paul Chervinsky, MD, North Dartmouth, MA; Jonathan Corren, MD, Los Angeles, CA; David Coutin, MD, Bend, OR; David Denmead, MD, Danville, CA; James Fish, MD, Philadelphia, PA; Charles Fogarty, MD, Spartanburg, SC; Pinkus Goldberg, MD, Indianapolis, IN; Marshall P. Grodofsky, MD, West Hartford, CT; Jay Grossman, MD, Tucson, AZ; Alan Heller, MD, San Jose, CA; Harold Kaiser, MD, Minneapolis, MN; Edward Kent, MD, South Burlington, VT; Kathy Lampl, MD, Rockville, MD; Allen Lieberman, MD, Austin, TX; Richard Lockey, MD, Tampa, FL; Stephen McGeady, MD, Philadelphia, PA; Jon Musmand, MD, Portland, ME; Anjuli Nayak, MD, Normal, IL; Gregory Neagos, MD, Ypsilanti, MI; John Oppenheimer, MD, Springfield, NJ; Frank Picone, MD, Tinton Falls, NJ; Jacob Pinnas, MD, Tucson, AZ; Gordon Raphael, MD, Bethesda, MD; Guy Settipane, MD, Providence, RI; William Sokol, MD, Newport Beach, CA; James Taylor, MD, Tacoma, WA; Robert Townley, MD, Omaha, NE. We also acknowledge the assistance of Cheryl Beale, MA, and Richard A. Rogers, MS, in the preparation and editing of this manuscript. This study was supported by a grant from GlaxoWellcome Inc., Research Triangle Park, North Carolina.

Asthma, a chronic inflammatory disorder of the airways, affects approximately 17.3 million people in the United States, with more than 5000 deaths per year attributed to the disorder.¹ National Institutes of Health (NIH) guidelines for asthma management recommend that patients with persistent asthma (ie, patients having symptoms more than twice weekly) receive daily medication for long-term control of asthma.¹ These guidelines consider inhaled corticosteroids the most effective controller medication for persistent asthma.

Recently, leukotriene modifiers have been introduced as a new class of asthma medications. NIH guidelines include leukotriene modifiers as potential alternative controller medications for mild, persistent asthma.¹ However, the guidelines also state that additional data are needed to establish the role and position of leukotriene modifiers in asthma therapy. To date, there have been few published reports of placebo-controlled clinical trials comparing the efficacy, safety, and quality of life in asthma patients treated with an inhaled corticosteroid versus a leukotriene modifier.

This study assessed the clinical benefits of an inhaled corticosteroid and a leukotriene modifier as first-line treatment for persistent asthma in patients who were symptomatic when using short-acting b2-agonists alone.

Methods

Patients

Patients aged 12 years old or older with asthma² were eligible if they had used a short-acting b2-agonist (either scheduled or as-needed) for at least 6 weeks preceding the study, if they had a 1-second forced expiratory volume (FEV₁) between 50% to 80% of predicted values, ^3-5 and if they demonstrated reversibility of FEV₁ (Ž12% increase within 30 minutes after inhaling 180 mg albuterol). Patients were excluded from the study if they had life-threatening asthma, significant and uncontrolled diseases (eg, chronic obstructive pulmonary disease, diabetes, coronary disease), used tobacco products within the preceding year, or had a smoking history of more than 10 pack-years. Patients were also excluded if they received any systemic corticosteroid within 6 months of screening, any inhaled corticosteroid within 1 month of screening, or any leukotriene modifier within 1 week of screening.

Study Design

This randomized double-blind, double-dummy, parallel-group study was conducted at 34 sites in the United States. All patients (or their parent or guardian) gave written informed consent before entering the study. The study was approved by institutional review boards for each site.

Eligible patients entered an 8-to 14-day run-in period to establish baseline respiratory function. Patients who met the study criteria were randomly assigned to receive 1 of the following double-blind, double-dummy treatments every morning and evening for 12 weeks: (1) 88 mg fluticasone propionate (FP [Flovent Inhalation Aerosol], 2 puffs of 44 mg each) administered by metered-dose inhaler (MDI) and a placebo capsule; (2) oral zafirlukast 20 mg (Accolate over-encapsulated tablet) and 2 puffs of placebo by MDI; or (3) a placebo capsule and 2 puffs of placebo by MDI. The zafirlukast dosage (20 mg twice daily) is the currently recommended dosage for patients aged 12 years or older, and 88 mg FP twice daily is the lowest FP dosage delivered by MDI approved for use in adolescents and adults. Treatment assignments were generated by computer in blocks of 6 so that each of the 3 treatments was represented twice in random order.

Before initiating the study, comparative dissolution testing was performed to determine whether over-encapsulation would affect the bioequivalence of the blinded zafirlukast tablets. The dissolution profiles were the same for the over-encapsulated (blinded drug) and the trade product. Empty capsules were also tested, and complete dissolution and disintegration of the capsules occurred within 5 minutes.

Patients were instructed to inhale 2 puffs from their study MDI each morning and evening and to swallow 1 study capsule each morning at least 1 hour before breakfast, and each evening at least 2 hours following dinner. The timing of administration of study capsules was based on the limited absorption of zafirlukast when given with food. Patients were instructed to rinse their mouths with water after using the MDI. Patients were not allowed to take any asthma medication during the study other than the study medication, albuterol (Ventolin Inhalation Aerosol) as needed for symptom relief, or oral or parenteral corticosteroids for asthma exacerbations. The use of medications that interact with inhaled corticosteroids or leukotriene modifiers was not allowed. Patients who experienced 2 exacerbations (eg, required a second burst of oral corticosteroids) or received oral or parenteral corticosteroids for more than 14 consecutive days were withdrawn from the study.

Assessments of Efficacy and Safety

The primary physiologic efficacy measure was morning predose FEV₁. Other efficacy measures included patient-recorded morning (am) and evening (pm) peak expiratory flow (PEF). Patient-oriented outcomes included asthma symptom scores, percentage of symptom-free days, albuterol use, percentage of albuterol-free days, and nighttime awakenings due to asthma symptoms; physicians’ overall assessment of medication effectiveness, and patient questionnaires concerning asthma-specific quality of life, productivity, and satisfaction with study medication. Safety assessments included the incidences of asthma exacerbations, adverse events, and pre-defined, clinically significant, laboratory test results.

FEV₁ was assessed at each clinic visit between 6 am and 10 am.⁶ Morning doses of study medications were withheld before each clinic visit, and the visit was rescheduled if the patient used albuterol within 6 hours preceding the visit. Before taking study medications, patients recorded on a diary card any asthma symptoms experienced, using a 6-point scale (where 0 = no symptoms and 5 = symptoms caused severe discomfort and prevented normal activities). After recording symptoms, patients measured and recorded their PEF, along with the number of puffs of albuterol used and the number of nighttime awakenings due to asthma. Patients reported study medication compliance daily on the diary card. Adverse events, concurrent medications, and diary cards were reviewed at each clinic visit. Laboratory tests were performed at the beginning of the run-in period and at the patient’s last visit.

At the last clinic visit, following a review of diary cards and the status of each patient, physicians rated the overall effectiveness of each patient’s blinded study medication using a 4-point scale (where 0 = ineffective and 3 = very effective). At weeks 2 and 12, patients assessed their overall satisfaction with study medication using a 7-point scale (where 0= very dissatisfied and 6 = very satisfied) and completed the Asthma Quality of Life Questionnaire (AQLQ),⁷ which measures the impact of asthma on patients’ daily functioning. At weeks 2, 4, 6, 8, and 12, patients documented their productivity (days missed from work or school and days with symptoms at work or school) over the previous 2 weeks.

Statistical Analyses

The sample size (Ž100 patients completing each treatment) provided Ž80% power to detect a difference in FEV₁ of 0.25 L between treatment groups, according to a 2-sample t test with a significance level of 0.05 and a standard deviation of 0.55 L.

All patients randomly assigned to the study drug were included in the safety analyses. Nine patients at 1 site were excluded from the efficacy analyses because of significant deviations from good clinical practice standards.

Change from baseline was calculated at each clinic visit and at the last available observation for FEV₁. For each subject-recorded assessment from the diary card, weekly means were calculated for each patient, and change from baseline (baseline = mean of last 7 days’ data from run-in) was calculated at each week and at the endpoint (endpoint = mean of last available weekly data). Onset of effect was evaluated as change in daily am PEF from day of randomization. Changes from baseline in global AQLQ and the individual domain scores were computed at the endpoint (endpoint = last available questionnaire). Analyses of covariance adjusted for investigator effect and included the baseline value as a covariate. Analysis of covariance F tests were used to evaluate treatment differences in all change from baseline values.

Cochran-Mantel-Haenszel tests, controlling for investigator, were used to determine treatment differences in physicians’ overall assessment of efficacy (at endpoint), in patient satisfaction with study medication (at week 2 and endpoint), and in the number of patients in each treatment group with at least 1 asthma exacerbation requiring oral or parenteral corticosteroids. Van Elteren tests were used to assess patient productivity.

Results

Patients

A total of 338 patients, 12 to 75 years of age, were randomly assigned to treatment (113, FP; 111, zafirlukast; 114, placebo). Equal numbers of men and women were enrolled in the study, and the majority of patients were non-Hispanic white (86%) or African American (10%). Baseline pulmonary function (% of predicted FEV₁) was similar across treatment groups (66%-67% predicted). The majority of patients in each treatment group had moderate asthma (77%, FP; 82%, zafirlukast; 18%, placebo), and most patients in each treatment group had asthma diagnosed for at least 10 years (64%-73%). A post hoc survey of patients revealed that the majority of respondents (149 of 286 patients; 52%) were recruited from primary care practices for the purpose of the study or were treating their asthma with over-the-counter bronchodilators. Similar numbers of patients in each group completed the study (81%-86% of patients). The most common reasons for discontinuation were lack of efficacy (2%-5%) and withdrawn consent (3%-5%). Median compliance was 93% in each group for both inhaled and oral study medication.

Efficacy

Treatment with FP significantly improved pulmonary function more than zafirlukast or placebo. At the endpoint Table 1 and at individual points as early as day 4 ( Figure 1, mean improvements in FEV₁ and in am PF and pm PEF were significantly greater in the FP group compared with the zafirlukast or placebo group (P <.05). When stratified by baseline pulmonary function (% predicted FEV₁), mean changes in FEV₁ and pm PEF were significantly greater in FP patients than in zafirlukast or placebo patients, regardless of baseline pulmonary function Table 2.

At the endpoint, treatment with FP significantly improved all clinical parameters (mean symptom scores, percentages of symptom-free and albuterol-free days, albuterol use, and nighttime awakenings due to asthma) compared with placebo (P≤.006, Table 1. In contrast, treatment with zafirlukast resulted in significant improvements in albuterol use and the percentage of symptom-free and albuterol-free days (P≤.040, vs placebo), but not in mean symptom scores or the number of nighttime awakenings. Compared with zafirlukast at the endpoint, FP produced significantly greater improvements in mean symptom scores, the percentages of symptom-free and albuterol-free days, albuterol use, and the number of nighttime awakenings due to asthma (P≤.038). When compared with placebo at individual time points, FP produced significantly greater improvements in each clinical parameter at most or all time points (P≤.039). In contrast, treatment with zafirlukast resulted in fewer time points with significantly different improvements in these parameters compared with placebo. Compared with zafirlukast, FP produced significantly greater improvements in mean symptom scores, the percentages of symptom-free and albuterol-free days, albuterol use, and nighttime awakenings due to asthma at the majority of time points from week 2 through week 12 (P≤.050).

At the endpoint, physicians’ overall assessments of study medication efficacy and patients’ overall satisfaction with study medication both significantly favored FP over zafirlukast (P≤.025) or placebo (P <.001; Figure 2. More physicians rated FP “effective” or “very effective” compared with zafirlukast or placebo. Likewise, more FP patients were “satisfied” or “very satisfied” with their study medication than were patients taking zafirlukast or placebo. The patient productivity questionnaire data showed that the mean number of days that patients attended work or school with asthma symptoms was significantly higher in the zafirlukast and placebo groups than the FP group (3.8, 4.4, and 1.8 days, respectively, P≤.03).

At the endpoint, treatment with FP produced significantly greater improvements in global and individual domain AQLQ scores compared with zafirlukast or placebo (P≤.033; Figure 3. Furthermore, treatment differences between FP and placebo for global and most individual domain AQLQ scores are considered clinically meaningful (Ž0.5).⁸ In contrast, zafirlukast did not achieve clinically meaningful differences compared with placebo for any AQLQ endpoint.

Safety

The percentages of patients who experienced at least 1 adverse event were similar across treatment groups (67% to 72%). More FP patients reported sinusitis (12%, vs 4% with zafirlukast and 4% with placebo), and more zafirlukast patients reported chest congestion (5%, vs <1% with FP and 0 with placebo). Oropharyngeal candidiasis was reported for 2 placebo recipients and 3 FP patients. Few patients experienced adverse events potentially related to study medication (12% to 13% of each group). The most common adverse events considered potentially related to study medication were throat irritation (FP: 4% of patients; zafirlukast and placebo: 3% each) and headache (FP: 3%; placebo and zafirlukast: 2% each). Two patients in the FP group and 1 patient in the zafirlukast group and 1 patient in the placebo group discontinued study treatment because of adverse events possibly related to study medication (FP: diarrhea, coughing and hoarseness; zafirlukast: increase in g-glutamyltransferase [GGT; other hepatic enzyme levels were not elevated]; placebo: lightheadedness and shakiness). The only adverse events requiring hospitalization occurred in the placebo group. Few clinically significant laboratory abnormalities ( 3% for any assay) were reported; none were considered related to study medication. Fewer FP patients experienced asthma exacerbations that required treatment with oral corticosteroids compared with zafirlukast or placebo patients (4%, 12%, and 10% of patients, respectively); treatment differences, however, were not statistically significant.

Discussion

The goals of asthma therapy are to prevent asthma symptoms, maintain normal activity levels, normalize lung function, meet patients’ expectations and satisfaction, and provide optimal therapy with minimal adverse effects.¹ For patients with persistent asthma, daily controller medications are recommended.¹ The results of our study indicate that, compared with zafirlukast or placebo, treatment with FP resulted in greater improvements in patients’ symptoms, quality of life, and overall satisfaction, as well as less albuterol use, fewer nighttime awakenings, and better pulmonary function. Furthermore, fewer FP patients experienced an asthma exacerbation compared with zafirlukast or placebo patients. Improvements in the zafirlukast group were generally similar to improvements reported in previous placebo-controlled trials with zafirlukast.^9,10

In our study, differences between FP and zafirlukast were seen early, and those differences generally persisted throughout the 12-week treatment. Treatment effects that favored FP were seen for all clinical endpoints by week 2, while differences in pulmonary function that favored FP were noted by day 4 of treatment. The rapid improvements noted with FP in our study are consistent with previously published results.¹¹ In the current study, patients treated with FP continued to improve with each subsequent week of treatment, while improvements with zafirlukast generally peaked after 2 weeks of treatment. Although the percentage of symptom-free days in the zafirlukast group appeared to increase in the last 4 weeks of the study, this may be related to the higher withdrawal rate during this period. Patient-rated outcomes support the clinical findings. Significant differences between FP and zafirlukast occurred by week 2 with quality of life, patient satisfaction, and days worked with asthma symptoms. The treatment effects noted in this study may be related to the rapid onset of action of fluticasone¹¹ and may result in improved compliance and better long-term asthma control.

Preventing recurrent exacerbations and minimizing the need for emergency department visits or hospitalizations is an important goal in asthma therapy.¹ The low incidence of exacerbations in the FP group in our study is consistent with data from previous clinical trials.^12-14 With regard to the frequency of asthma exacerbations in patients receiving leukotriene modifiers, available data demonstrate that zileuton, zafirlukast, and montelukast each significantly decrease the risk of asthma exacerbations compared with placebo.^10,15-17 Although leukotriene modifiers are superior to placebo in reducing the risk of asthma exacerbations, the results of our study indicate that patients receiving an inhaled corticosteroid had fewer asthma exacerbations than those receiving a leukotriene receptor antagonist. These data are in agreement with an earlier placebo-controlled, comparative study.¹⁷

A significant placebo effect was seen with FEV₁ in our study and is similar to observations from other FP studies.^18,19 In our study, the placebo effect was larger in patients with a percent predicted FEV¹ of more than 70% at baseline. One possible explanation for this is regression to the mean, since patients typically present for care when their symptoms are worst. Another is that milder patients are affected more by the additional medical attention received in clinical studies or that participation in a clinical study influenced patients to avoid asthma-triggering events. The placebo effect was strongest with the spirometric endpoints, for reasons that are unclear. Regardless of the magnitude of the placebo response, the differences in our study between placebo and each active treatment are consistent with the differences seen in other studies involving FP or zafirlukast.^18,20

Because asthma involves chronic inflammation of the airways, the NIH guidelines recommend long-term controller medications with anti-inflammatory effects. In the present study, an inhaled cortico-steroid and a leukotriene modifier both led to improvements in pulmonary function. However, FP continued to improve pulmonary function throughout the 12-week study, while the benefits of zafirlukast tended to plateau by week 2 of treatment. Although the precise mechanism of action of glucocorticoids in asthma is unknown, the greater degree of effectiveness seen with FP in the present study may be related to the multiple effects of inhaled corticosteroids on airway inflammation.²¹ A post hoc analysis of the patient population stratified by baseline pulmonary function (% predicted FEV₁ of 50% to 70% or 70.1% to 80%) showed that, regardless of baseline lung function, FP was more effective than zafirlukast or placebo with regard to improvement in albuterol use and pulmonary function. These data are consistent with asthma guidelines, which recommend inhaled corticosteroids as first-line treatment for all levels of persistent asthma.¹

Limitations

The design of our study has several potential limitations. First, several studies indicate that dosages of zafirlukast higher than 20 mg twice daily may result in greater pharmacologic results.^22-26 Our study, however, was performed in the United States, where this is the only approved dosage for zafirlukast. Second, outcomes were measured at the end of the dosing interval, before administration of morning dosages of study medication. The timing of these tests may have introduced a potential bias in favor of FP, since there is evidence to support bronchodilatory activity with leukotriene modifiers.²⁷ The lesser effect of zafirlukast could be explained if the drug has a duration of action that is shorter than the recommended 12-hour dosing interval. However, the procedures and timing used in this study are similar to those used in other studies of leukotriene modifiers.^9,10,17 Third, a large number of efficacy measures were collected at multiple time points, which potentially increases the likelihood of finding significant treatment differences. However, many of the differences noted in this study were highly significant, and all of the measures were in agreement. This suggests that treatment differences seen in this study were valid. Finally, zafirlukast tablets were over-encapsulated for the purpose of blinding. The over-encapsulation did not appear to affect the performance of zafirlukast, because over-encapsulated and nonencapsulated tablets exhibited similar dissolution profiles. This conclusion is reinforced by the fact that the efficacy profile of zafirlukast in our study is similar to that seen in previous studies.^9,10

Conclusions

Treatment with FP 88 mg twice daily was more effective than zafirlukast 20 mg twice daily or placebo in improving asthma control and patient-rated outcomes in patients with persistent asthma who were previously receiving short-acting b2-agonists alone.

Acknowledgments

We acknowledge the important contributions made by the clinical investigators in this study: Thomas Bell, MD, Missoula, MT; Robert Berkowitz, MD, Atlanta, GA; David Bernstein, MD, Cincinnati, OH; Paul Chervinsky, MD, North Dartmouth, MA; Jonathan Corren, MD, Los Angeles, CA; David Coutin, MD, Bend, OR; David Denmead, MD, Danville, CA; James Fish, MD, Philadelphia, PA; Charles Fogarty, MD, Spartanburg, SC; Pinkus Goldberg, MD, Indianapolis, IN; Marshall P. Grodofsky, MD, West Hartford, CT; Jay Grossman, MD, Tucson, AZ; Alan Heller, MD, San Jose, CA; Harold Kaiser, MD, Minneapolis, MN; Edward Kent, MD, South Burlington, VT; Kathy Lampl, MD, Rockville, MD; Allen Lieberman, MD, Austin, TX; Richard Lockey, MD, Tampa, FL; Stephen McGeady, MD, Philadelphia, PA; Jon Musmand, MD, Portland, ME; Anjuli Nayak, MD, Normal, IL; Gregory Neagos, MD, Ypsilanti, MI; John Oppenheimer, MD, Springfield, NJ; Frank Picone, MD, Tinton Falls, NJ; Jacob Pinnas, MD, Tucson, AZ; Gordon Raphael, MD, Bethesda, MD; Guy Settipane, MD, Providence, RI; William Sokol, MD, Newport Beach, CA; James Taylor, MD, Tacoma, WA; Robert Townley, MD, Omaha, NE. We also acknowledge the assistance of Cheryl Beale, MA, and Richard A. Rogers, MS, in the preparation and editing of this manuscript. This study was supported by a grant from GlaxoWellcome Inc., Research Triangle Park, North Carolina.

Asthma, a chronic inflammatory disorder of the airways, affects approximately 17.3 million people in the United States, with more than 5000 deaths per year attributed to the disorder.¹ National Institutes of Health (NIH) guidelines for asthma management recommend that patients with persistent asthma (ie, patients having symptoms more than twice weekly) receive daily medication for long-term control of asthma.¹ These guidelines consider inhaled corticosteroids the most effective controller medication for persistent asthma.

Recently, leukotriene modifiers have been introduced as a new class of asthma medications. NIH guidelines include leukotriene modifiers as potential alternative controller medications for mild, persistent asthma.¹ However, the guidelines also state that additional data are needed to establish the role and position of leukotriene modifiers in asthma therapy. To date, there have been few published reports of placebo-controlled clinical trials comparing the efficacy, safety, and quality of life in asthma patients treated with an inhaled corticosteroid versus a leukotriene modifier.

This study assessed the clinical benefits of an inhaled corticosteroid and a leukotriene modifier as first-line treatment for persistent asthma in patients who were symptomatic when using short-acting b2-agonists alone.

Methods

Patients

Patients aged 12 years old or older with asthma² were eligible if they had used a short-acting b2-agonist (either scheduled or as-needed) for at least 6 weeks preceding the study, if they had a 1-second forced expiratory volume (FEV₁) between 50% to 80% of predicted values, ^3-5 and if they demonstrated reversibility of FEV₁ (Ž12% increase within 30 minutes after inhaling 180 mg albuterol). Patients were excluded from the study if they had life-threatening asthma, significant and uncontrolled diseases (eg, chronic obstructive pulmonary disease, diabetes, coronary disease), used tobacco products within the preceding year, or had a smoking history of more than 10 pack-years. Patients were also excluded if they received any systemic corticosteroid within 6 months of screening, any inhaled corticosteroid within 1 month of screening, or any leukotriene modifier within 1 week of screening.

Study Design

This randomized double-blind, double-dummy, parallel-group study was conducted at 34 sites in the United States. All patients (or their parent or guardian) gave written informed consent before entering the study. The study was approved by institutional review boards for each site.

Eligible patients entered an 8-to 14-day run-in period to establish baseline respiratory function. Patients who met the study criteria were randomly assigned to receive 1 of the following double-blind, double-dummy treatments every morning and evening for 12 weeks: (1) 88 mg fluticasone propionate (FP [Flovent Inhalation Aerosol], 2 puffs of 44 mg each) administered by metered-dose inhaler (MDI) and a placebo capsule; (2) oral zafirlukast 20 mg (Accolate over-encapsulated tablet) and 2 puffs of placebo by MDI; or (3) a placebo capsule and 2 puffs of placebo by MDI. The zafirlukast dosage (20 mg twice daily) is the currently recommended dosage for patients aged 12 years or older, and 88 mg FP twice daily is the lowest FP dosage delivered by MDI approved for use in adolescents and adults. Treatment assignments were generated by computer in blocks of 6 so that each of the 3 treatments was represented twice in random order.

Before initiating the study, comparative dissolution testing was performed to determine whether over-encapsulation would affect the bioequivalence of the blinded zafirlukast tablets. The dissolution profiles were the same for the over-encapsulated (blinded drug) and the trade product. Empty capsules were also tested, and complete dissolution and disintegration of the capsules occurred within 5 minutes.

Patients were instructed to inhale 2 puffs from their study MDI each morning and evening and to swallow 1 study capsule each morning at least 1 hour before breakfast, and each evening at least 2 hours following dinner. The timing of administration of study capsules was based on the limited absorption of zafirlukast when given with food. Patients were instructed to rinse their mouths with water after using the MDI. Patients were not allowed to take any asthma medication during the study other than the study medication, albuterol (Ventolin Inhalation Aerosol) as needed for symptom relief, or oral or parenteral corticosteroids for asthma exacerbations. The use of medications that interact with inhaled corticosteroids or leukotriene modifiers was not allowed. Patients who experienced 2 exacerbations (eg, required a second burst of oral corticosteroids) or received oral or parenteral corticosteroids for more than 14 consecutive days were withdrawn from the study.

Assessments of Efficacy and Safety

The primary physiologic efficacy measure was morning predose FEV₁. Other efficacy measures included patient-recorded morning (am) and evening (pm) peak expiratory flow (PEF). Patient-oriented outcomes included asthma symptom scores, percentage of symptom-free days, albuterol use, percentage of albuterol-free days, and nighttime awakenings due to asthma symptoms; physicians’ overall assessment of medication effectiveness, and patient questionnaires concerning asthma-specific quality of life, productivity, and satisfaction with study medication. Safety assessments included the incidences of asthma exacerbations, adverse events, and pre-defined, clinically significant, laboratory test results.

FEV₁ was assessed at each clinic visit between 6 am and 10 am.⁶ Morning doses of study medications were withheld before each clinic visit, and the visit was rescheduled if the patient used albuterol within 6 hours preceding the visit. Before taking study medications, patients recorded on a diary card any asthma symptoms experienced, using a 6-point scale (where 0 = no symptoms and 5 = symptoms caused severe discomfort and prevented normal activities). After recording symptoms, patients measured and recorded their PEF, along with the number of puffs of albuterol used and the number of nighttime awakenings due to asthma. Patients reported study medication compliance daily on the diary card. Adverse events, concurrent medications, and diary cards were reviewed at each clinic visit. Laboratory tests were performed at the beginning of the run-in period and at the patient’s last visit.

At the last clinic visit, following a review of diary cards and the status of each patient, physicians rated the overall effectiveness of each patient’s blinded study medication using a 4-point scale (where 0 = ineffective and 3 = very effective). At weeks 2 and 12, patients assessed their overall satisfaction with study medication using a 7-point scale (where 0= very dissatisfied and 6 = very satisfied) and completed the Asthma Quality of Life Questionnaire (AQLQ),⁷ which measures the impact of asthma on patients’ daily functioning. At weeks 2, 4, 6, 8, and 12, patients documented their productivity (days missed from work or school and days with symptoms at work or school) over the previous 2 weeks.

Statistical Analyses

The sample size (Ž100 patients completing each treatment) provided Ž80% power to detect a difference in FEV₁ of 0.25 L between treatment groups, according to a 2-sample t test with a significance level of 0.05 and a standard deviation of 0.55 L.

All patients randomly assigned to the study drug were included in the safety analyses. Nine patients at 1 site were excluded from the efficacy analyses because of significant deviations from good clinical practice standards.

Change from baseline was calculated at each clinic visit and at the last available observation for FEV₁. For each subject-recorded assessment from the diary card, weekly means were calculated for each patient, and change from baseline (baseline = mean of last 7 days’ data from run-in) was calculated at each week and at the endpoint (endpoint = mean of last available weekly data). Onset of effect was evaluated as change in daily am PEF from day of randomization. Changes from baseline in global AQLQ and the individual domain scores were computed at the endpoint (endpoint = last available questionnaire). Analyses of covariance adjusted for investigator effect and included the baseline value as a covariate. Analysis of covariance F tests were used to evaluate treatment differences in all change from baseline values.

Cochran-Mantel-Haenszel tests, controlling for investigator, were used to determine treatment differences in physicians’ overall assessment of efficacy (at endpoint), in patient satisfaction with study medication (at week 2 and endpoint), and in the number of patients in each treatment group with at least 1 asthma exacerbation requiring oral or parenteral corticosteroids. Van Elteren tests were used to assess patient productivity.

Results

Patients

A total of 338 patients, 12 to 75 years of age, were randomly assigned to treatment (113, FP; 111, zafirlukast; 114, placebo). Equal numbers of men and women were enrolled in the study, and the majority of patients were non-Hispanic white (86%) or African American (10%). Baseline pulmonary function (% of predicted FEV₁) was similar across treatment groups (66%-67% predicted). The majority of patients in each treatment group had moderate asthma (77%, FP; 82%, zafirlukast; 18%, placebo), and most patients in each treatment group had asthma diagnosed for at least 10 years (64%-73%). A post hoc survey of patients revealed that the majority of respondents (149 of 286 patients; 52%) were recruited from primary care practices for the purpose of the study or were treating their asthma with over-the-counter bronchodilators. Similar numbers of patients in each group completed the study (81%-86% of patients). The most common reasons for discontinuation were lack of efficacy (2%-5%) and withdrawn consent (3%-5%). Median compliance was 93% in each group for both inhaled and oral study medication.

Efficacy

Treatment with FP significantly improved pulmonary function more than zafirlukast or placebo. At the endpoint Table 1 and at individual points as early as day 4 ( Figure 1, mean improvements in FEV₁ and in am PF and pm PEF were significantly greater in the FP group compared with the zafirlukast or placebo group (P <.05). When stratified by baseline pulmonary function (% predicted FEV₁), mean changes in FEV₁ and pm PEF were significantly greater in FP patients than in zafirlukast or placebo patients, regardless of baseline pulmonary function Table 2.

At the endpoint, treatment with FP significantly improved all clinical parameters (mean symptom scores, percentages of symptom-free and albuterol-free days, albuterol use, and nighttime awakenings due to asthma) compared with placebo (P≤.006, Table 1. In contrast, treatment with zafirlukast resulted in significant improvements in albuterol use and the percentage of symptom-free and albuterol-free days (P≤.040, vs placebo), but not in mean symptom scores or the number of nighttime awakenings. Compared with zafirlukast at the endpoint, FP produced significantly greater improvements in mean symptom scores, the percentages of symptom-free and albuterol-free days, albuterol use, and the number of nighttime awakenings due to asthma (P≤.038). When compared with placebo at individual time points, FP produced significantly greater improvements in each clinical parameter at most or all time points (P≤.039). In contrast, treatment with zafirlukast resulted in fewer time points with significantly different improvements in these parameters compared with placebo. Compared with zafirlukast, FP produced significantly greater improvements in mean symptom scores, the percentages of symptom-free and albuterol-free days, albuterol use, and nighttime awakenings due to asthma at the majority of time points from week 2 through week 12 (P≤.050).

At the endpoint, physicians’ overall assessments of study medication efficacy and patients’ overall satisfaction with study medication both significantly favored FP over zafirlukast (P≤.025) or placebo (P <.001; Figure 2. More physicians rated FP “effective” or “very effective” compared with zafirlukast or placebo. Likewise, more FP patients were “satisfied” or “very satisfied” with their study medication than were patients taking zafirlukast or placebo. The patient productivity questionnaire data showed that the mean number of days that patients attended work or school with asthma symptoms was significantly higher in the zafirlukast and placebo groups than the FP group (3.8, 4.4, and 1.8 days, respectively, P≤.03).

At the endpoint, treatment with FP produced significantly greater improvements in global and individual domain AQLQ scores compared with zafirlukast or placebo (P≤.033; Figure 3. Furthermore, treatment differences between FP and placebo for global and most individual domain AQLQ scores are considered clinically meaningful (Ž0.5).⁸ In contrast, zafirlukast did not achieve clinically meaningful differences compared with placebo for any AQLQ endpoint.

Safety

The percentages of patients who experienced at least 1 adverse event were similar across treatment groups (67% to 72%). More FP patients reported sinusitis (12%, vs 4% with zafirlukast and 4% with placebo), and more zafirlukast patients reported chest congestion (5%, vs <1% with FP and 0 with placebo). Oropharyngeal candidiasis was reported for 2 placebo recipients and 3 FP patients. Few patients experienced adverse events potentially related to study medication (12% to 13% of each group). The most common adverse events considered potentially related to study medication were throat irritation (FP: 4% of patients; zafirlukast and placebo: 3% each) and headache (FP: 3%; placebo and zafirlukast: 2% each). Two patients in the FP group and 1 patient in the zafirlukast group and 1 patient in the placebo group discontinued study treatment because of adverse events possibly related to study medication (FP: diarrhea, coughing and hoarseness; zafirlukast: increase in g-glutamyltransferase [GGT; other hepatic enzyme levels were not elevated]; placebo: lightheadedness and shakiness). The only adverse events requiring hospitalization occurred in the placebo group. Few clinically significant laboratory abnormalities ( 3% for any assay) were reported; none were considered related to study medication. Fewer FP patients experienced asthma exacerbations that required treatment with oral corticosteroids compared with zafirlukast or placebo patients (4%, 12%, and 10% of patients, respectively); treatment differences, however, were not statistically significant.

Discussion

The goals of asthma therapy are to prevent asthma symptoms, maintain normal activity levels, normalize lung function, meet patients’ expectations and satisfaction, and provide optimal therapy with minimal adverse effects.¹ For patients with persistent asthma, daily controller medications are recommended.¹ The results of our study indicate that, compared with zafirlukast or placebo, treatment with FP resulted in greater improvements in patients’ symptoms, quality of life, and overall satisfaction, as well as less albuterol use, fewer nighttime awakenings, and better pulmonary function. Furthermore, fewer FP patients experienced an asthma exacerbation compared with zafirlukast or placebo patients. Improvements in the zafirlukast group were generally similar to improvements reported in previous placebo-controlled trials with zafirlukast.^9,10

In our study, differences between FP and zafirlukast were seen early, and those differences generally persisted throughout the 12-week treatment. Treatment effects that favored FP were seen for all clinical endpoints by week 2, while differences in pulmonary function that favored FP were noted by day 4 of treatment. The rapid improvements noted with FP in our study are consistent with previously published results.¹¹ In the current study, patients treated with FP continued to improve with each subsequent week of treatment, while improvements with zafirlukast generally peaked after 2 weeks of treatment. Although the percentage of symptom-free days in the zafirlukast group appeared to increase in the last 4 weeks of the study, this may be related to the higher withdrawal rate during this period. Patient-rated outcomes support the clinical findings. Significant differences between FP and zafirlukast occurred by week 2 with quality of life, patient satisfaction, and days worked with asthma symptoms. The treatment effects noted in this study may be related to the rapid onset of action of fluticasone¹¹ and may result in improved compliance and better long-term asthma control.

Preventing recurrent exacerbations and minimizing the need for emergency department visits or hospitalizations is an important goal in asthma therapy.¹ The low incidence of exacerbations in the FP group in our study is consistent with data from previous clinical trials.^12-14 With regard to the frequency of asthma exacerbations in patients receiving leukotriene modifiers, available data demonstrate that zileuton, zafirlukast, and montelukast each significantly decrease the risk of asthma exacerbations compared with placebo.^10,15-17 Although leukotriene modifiers are superior to placebo in reducing the risk of asthma exacerbations, the results of our study indicate that patients receiving an inhaled corticosteroid had fewer asthma exacerbations than those receiving a leukotriene receptor antagonist. These data are in agreement with an earlier placebo-controlled, comparative study.¹⁷

A significant placebo effect was seen with FEV₁ in our study and is similar to observations from other FP studies.^18,19 In our study, the placebo effect was larger in patients with a percent predicted FEV¹ of more than 70% at baseline. One possible explanation for this is regression to the mean, since patients typically present for care when their symptoms are worst. Another is that milder patients are affected more by the additional medical attention received in clinical studies or that participation in a clinical study influenced patients to avoid asthma-triggering events. The placebo effect was strongest with the spirometric endpoints, for reasons that are unclear. Regardless of the magnitude of the placebo response, the differences in our study between placebo and each active treatment are consistent with the differences seen in other studies involving FP or zafirlukast.^18,20

Because asthma involves chronic inflammation of the airways, the NIH guidelines recommend long-term controller medications with anti-inflammatory effects. In the present study, an inhaled cortico-steroid and a leukotriene modifier both led to improvements in pulmonary function. However, FP continued to improve pulmonary function throughout the 12-week study, while the benefits of zafirlukast tended to plateau by week 2 of treatment. Although the precise mechanism of action of glucocorticoids in asthma is unknown, the greater degree of effectiveness seen with FP in the present study may be related to the multiple effects of inhaled corticosteroids on airway inflammation.²¹ A post hoc analysis of the patient population stratified by baseline pulmonary function (% predicted FEV₁ of 50% to 70% or 70.1% to 80%) showed that, regardless of baseline lung function, FP was more effective than zafirlukast or placebo with regard to improvement in albuterol use and pulmonary function. These data are consistent with asthma guidelines, which recommend inhaled corticosteroids as first-line treatment for all levels of persistent asthma.¹

Limitations

The design of our study has several potential limitations. First, several studies indicate that dosages of zafirlukast higher than 20 mg twice daily may result in greater pharmacologic results.^22-26 Our study, however, was performed in the United States, where this is the only approved dosage for zafirlukast. Second, outcomes were measured at the end of the dosing interval, before administration of morning dosages of study medication. The timing of these tests may have introduced a potential bias in favor of FP, since there is evidence to support bronchodilatory activity with leukotriene modifiers.²⁷ The lesser effect of zafirlukast could be explained if the drug has a duration of action that is shorter than the recommended 12-hour dosing interval. However, the procedures and timing used in this study are similar to those used in other studies of leukotriene modifiers.^9,10,17 Third, a large number of efficacy measures were collected at multiple time points, which potentially increases the likelihood of finding significant treatment differences. However, many of the differences noted in this study were highly significant, and all of the measures were in agreement. This suggests that treatment differences seen in this study were valid. Finally, zafirlukast tablets were over-encapsulated for the purpose of blinding. The over-encapsulation did not appear to affect the performance of zafirlukast, because over-encapsulated and nonencapsulated tablets exhibited similar dissolution profiles. This conclusion is reinforced by the fact that the efficacy profile of zafirlukast in our study is similar to that seen in previous studies.^9,10

Conclusions

Treatment with FP 88 mg twice daily was more effective than zafirlukast 20 mg twice daily or placebo in improving asthma control and patient-rated outcomes in patients with persistent asthma who were previously receiving short-acting b2-agonists alone.

Acknowledgments

We acknowledge the important contributions made by the clinical investigators in this study: Thomas Bell, MD, Missoula, MT; Robert Berkowitz, MD, Atlanta, GA; David Bernstein, MD, Cincinnati, OH; Paul Chervinsky, MD, North Dartmouth, MA; Jonathan Corren, MD, Los Angeles, CA; David Coutin, MD, Bend, OR; David Denmead, MD, Danville, CA; James Fish, MD, Philadelphia, PA; Charles Fogarty, MD, Spartanburg, SC; Pinkus Goldberg, MD, Indianapolis, IN; Marshall P. Grodofsky, MD, West Hartford, CT; Jay Grossman, MD, Tucson, AZ; Alan Heller, MD, San Jose, CA; Harold Kaiser, MD, Minneapolis, MN; Edward Kent, MD, South Burlington, VT; Kathy Lampl, MD, Rockville, MD; Allen Lieberman, MD, Austin, TX; Richard Lockey, MD, Tampa, FL; Stephen McGeady, MD, Philadelphia, PA; Jon Musmand, MD, Portland, ME; Anjuli Nayak, MD, Normal, IL; Gregory Neagos, MD, Ypsilanti, MI; John Oppenheimer, MD, Springfield, NJ; Frank Picone, MD, Tinton Falls, NJ; Jacob Pinnas, MD, Tucson, AZ; Gordon Raphael, MD, Bethesda, MD; Guy Settipane, MD, Providence, RI; William Sokol, MD, Newport Beach, CA; James Taylor, MD, Tacoma, WA; Robert Townley, MD, Omaha, NE. We also acknowledge the assistance of Cheryl Beale, MA, and Richard A. Rogers, MS, in the preparation and editing of this manuscript. This study was supported by a grant from GlaxoWellcome Inc., Research Triangle Park, North Carolina.

Millions of women with acute dysuria show up at offices, urgent care centers, and emergency departments for suspected urinary tract infections (UTIs), accounting for more than $1 billion in direct costs.¹ Since most UTIs are uncomplicated, numerous strategies have been proposed for managing them in more efficient and less costly ways. Berg² found 82 separate management strategies among 137 family physicians, with costs ranging from negligible to $250.

In a previous study³ we used a cost-utility analysis to demonstrate that among office-based approaches, empiric therapy of suspected UTIs was most cost-effective. This was robust over a wide range of assumptions, including prevalence, test characteristics, costs, complication rates, and outcomes. These findings were recently confirmed by Fenwick and colleagues⁴ in a British analysis. Among the approaches commonly used, telephone management has the potential for reducing cost, increasing convenience for patients, and reducing barriers to care. Saint and coworkers⁵ demonstrated that a clinical practice guideline using telephone-based management of presumed UTIs reduced the use of urine tests and increased the use of guideline-specified antibiotics. Although telephone management is common, we were unable to find any studies directly comparing that approach with office-based care.

We report on the first trial in which women with suspected UTIs were randomly assigned to telephone management or office-based management. The purpose of our study was to identify the prevalence of UTIs in women presenting with suspected UTIs, to determine if telephone management was comparable in outcomes with those of office-based care, and to determine if women whose infections were managed by telephone were satisfied with their care.

Methods

Setting

We enrolled patients at 6 primary care offices (including a walk-in clinic) of the Upper Peninsula Research Network (UPRNet) and the Department of Family Practice at Michigan State University in East Lansing. UPRNet is a primary care research network in the Upper Peninsula of Michigan.

Subjects

Nonpregnant women 18 years or older completed an interview to confirm eligibility when they called their primary care physicians with a suspected uncomplicated UTI. We defined a suspected uncomplicated UTI as a complaint of dysuria, pain on urination, complaint of urinary urgency and frequency, or the patient’s saying, “I think I have a bladder infection.” Patients with symptoms compatible with pyelonephritis (fever, chills, sweats, back or flank pain, or vomiting), vaginitis, or cervicitis (presence of a new or changed vaginal discharge) were excluded from our study. We also excluded women with diabetes, a previous history of pyelonephritis or other complicated UTI, a UTI in the preceding month, symptoms lasting longer than 14 days, and known kidney disease, anatomic abnormalities, or previous renal surgery. In addition, we excluded women receiving chemotherapy and those who had received antibiotics in the preceding month. Informed consent was obtained. Enrollment occurred between October of 1997 and March of 1999. All enrolled patients received $25 for participating in the study. The Michigan State University Committee on Research Involving Human Subjects approved our study.

Procedures and Measures

We stratified each practice, and using a central computer-generated random number, we blindly allocated patients to either treatment by telephone (telephone group) or office-based care (control group) by using an opaque envelope containing the group assignment. The envelope also provided instructions appropriate for the assigned group. We asked the telephone strategy patients to come to the clinic to leave a urine sample and pick up a prescription for an antibiotic; the control (usual care) patients were given a same-day appointment for a regular clinic visit.

We asked patients enrolled in both groups to rate the severity of urinary dysuria, urgency, frequency, function, and how they generally felt about their symptoms. Each was rated on a 10-point scale (10 was most severe). The telephone management patients were given a prescription for sulfamethoxazole/trimethoprim (800 mg/160 mg) twice daily for 7 days. If the patient was allergic to sulfa, she received a prescription for nitrofurantoin 100 mg twice daily for 7 days. Patients were required to have a urinalysis and urine culture before receiving the prescription. We asked the health care providers of the control patients to use their usual management strategies. The control group patients were also required to have a urinalysis and urine culture.

A nurse telephoned all women in both groups for follow-up 3 and 10 days after the time of enrollment. During these telephone calls we assessed patient status (including symptom scores and patient satisfaction) and whether the patient sought care with any practitioner since the initial time of enrollment. If UTI symptoms were still evident at either the 3- or 10-day follow-up, the patient was asked to make an appointment to see her practitioner. We recorded start and stop times for all interviews to assist in estimating and comparing costs of care for the 2 groups.

Primary Outcomes

The primary outcomes for our study were the UTI score and overall evaluation rating (OER)⁶ of the treatment experience. The UTI score is the sum of the ratings of severity of dysuria, urgency, frequency, function, and general symptoms (range=0-50). Since the UTI score has not been previously used, we pilot tested it on 20 women. These same women were reevaluated 2 to 3 days later. Test-retest reliability (measured by the overall correlation between the same items asked on 2 separate occasions with the order of questions changed) was 0.98. Validity (measured by the correlation coefficient between specific questions and a global question for the episode) was 0.52 for burning, 0.89 for frequency, 0.95 for urgency, 0.86 for interference with activities of daily living, and -0.76 for the actual interval between urination (smaller interval associated with greater overall episode severity).

The OER consisted of 2 questions about the overall quality of care and the outcome of care. Each question—answered as poor, fair, good, very good, or excellent—was scored from 1 to 5, respectively. The OER, a validated score,⁶ is a simple sum of the scores for these items (range 2-10, not normally distributed). Also, we asked the women about their overall satisfaction on a 5-point Likert scale.

Secondary Outcomes

We also evaluated urine culture results. We defined a negative culture as one with either no growth or less than 1000 colony-forming units. A positive culture had any growth of a single organism. A contaminated specimen (mixed flora) was rated as a negative culture, since this is how these are usually handled clinically.

Statistical Analysis

The data were entered into a database, and all entries were double-checked by one of the investigators for transcription errors. We analyzed the data using SAS software (version 7, SAS Institute, Cary, NC). Continuous variables (age, time variables, UTI score) were compared by treatment group using unpaired Student t tests. We used the Shapiro-Wilk test and the Kolmogorov-Smirnov test to confirm that the UTI score and changes in that score between assessments were normally distributed. We compared categorical variables (resolution of symptoms, culture results) by treatment group using a chi-square. We used Wilcoxon rank sum tests to compare ordinal data (OER) between the treatment groups.

For all sample size estimations we wanted to achieve 80% power with two-sided a of 0.05. To detect a 5-point difference in the UTI score, with an estimated a priori standard deviation (SD) of 5, we calculated that 16 subjects in each group are needed (21 for 90% power). With an SD of 7.5, 36 subjects in each group are needed. For patient satisfaction (OER), we estimated a priori that we would need 15 subjects in each group (19 for 90% power) to detect a 1-point difference with an estimated mean of 3.9 and a SD of 0.95.

Results

We identified 201 women with suspected UTIs of whom 99 were not eligible. Of the 102 eligible women, 30 declined to participate. We randomized 36 women to office care (control group) and 36 to telephone management. The subjects were predominantly white (n=62, 86%) and young (mean age=36.6 years, SD=12.3). Five patients had no culture results. Of the 67 remaining cultures, 4 (6%) were contaminated specimens, 20 (29.8%) had negative cultures, and 43 (64.2%) had positive cultures. Of the positive cultures, 34 (79.1%) grew Escherichia coli. Twenty-three of 34 (67.6%) cultures in the control group were positive compared with 20 of 33 (60.6%) in the telephone group (chi-square=0.3611; P=.55).

The groups were similar at baseline Table 1. On day 3 and day 10 there were no significant differences in the change in symptom scores or overall UTI score from baseline. We also found no difference in the change in urinary intervals from baseline. Table 2 shows these data. There was no significant difference in the overall evaluation rating. We also found no difference in satisfaction with care (median response was “very good” in the control group and “excellent” in the telephone group). These are shown in Table 3

On the third day after therapy was inititated, 20 of 33 (60.6%) of the control subjects had persistent urinary symptoms compared with 19 of 34 (55.8%, chi-square=0.1536; P=.70) in the telephone group. By day 10, 6 of 35 control patients (17.1%) had persistent symptoms, compared with 12 of 35 (34.3%) in the telephone group (chi-square=2.6923; P=0.1). Among the patients still symptomatic on the third day, culture results were available for 35, 11 (31%) of which were negative. Among those still symptomatic on the 10th day, 18 had available cultures, 9 (50%) of which were negative.

To evaluate the patients with persistent symptoms at the conclusion of the study we looked at the baseline and final UTI scores and baseline culture results. Six patients in the control group reported persistent symptoms. Three of these patients had final UTI scores less than 10, and 3 had UTI scores greater than 20. Two of these patients also had negative cultures. Twelve patients in the telephone group reported persistent symptoms; all but 1 had final UTI scores less than 10, and only 1 had a final UTI score greater than 20. Seven of the 12 patients had negative cultures at baseline.

We attempted to determine how office care differed from telephone care. Three patients in the control group received no antibiotics. Two of these had negative cultures, and no culture result was available for the third patient. All patients in the telephone group were prescribed antibiotics. Five control group patients who ultimately had positive cultures took antibiotics for less than 7 days, compared with only 3 in the telephone group. Among those receiving antibiotics, 30 of the control group patients received either sulfamethoxazole/trimethoprim or nitrofurantoin, and 3 received second-line agents. Because of allergies, 1 patient in the telephone group did not receive the planned therapy and received cephalexin instead.

We also evaluated the nursing time to administer various elements of the protocol. It took 2.5 minutes (SD=1.3) to determine eligibility to participate in the study and 5.3 minutes (SD=2.1) to enroll the subjects into the study. The nursing time for the day 3 follow-up took on average 5.6 minutes (SD=2.9) and 5.2 minutes (SD=2.0) on day 10.

Discussion

Although managing uncomplicated UTIs by telephone is a common practice in ambulatory primary care settings, we had no previous empiric evidence of its effectiveness compared with seeing patients in the office. In this randomized trial of office management versus management by telephone, two thirds of the women enrolled had culture-confirmed UTIs. The rate was similar in each group and mirrors that reported in the literature.⁷ We found no difference in improvement in symptom scores from baseline and no significant difference in overall satisfaction with the care provided or the outcome.

Gallagher and colleagues⁸ reported that when acute medical problems are triaged by nurses, patients are generally satisfied with care. However, UTIs represented only 5% of the telephone encounters. Delichatsios and coworkers⁹ similarly reported that patients calling to speak with the physician were generally satisfied with the advice given on the telephone, but they did not report outcomes related to specific conditions or therapies. Although 2 independent economic evaluations^3,4 have found empiric therapy to be cost-effective, neither included a strategy that avoids an office visit.

The direct cost of telephone management of uncomplicated UTIs is relatively low. It took only 2.5 minutes of nurse time to identify symptomatic women with risk factors for complicated UTIs who were good candidates for telephone management. This may cause a dilemma. Physicians practicing in predominantly fee-for-service settings will lose income by managing UTIs by telephone. In managed care settings, the financial incentives to reduce utilization make this practice inexpensive while simultaneously maintaining high patient satisfaction. Many physicians, however, complain about the complexity of the patients they now see, and having an occasional uncomplicated UTI might provide some breathing space on hectic days.

Limitations

We did not ask the practitioners who provided office-based care to alter their usual approach. By patient report, only 3 control patients received no antibiotics. This may reflect a knee-jerk response in which antibiotics are prescribed for all women with a suspected UTI. It may also reflect a very appropriate therapeutic threshold where physicians have a gestalt about the probability of a UTI that exceeds any diagnostic uncertainty. Although this has been described explicitly,¹⁰ we believe that seasoned clinicians do this implicitly. We did not attempt to open the “black box” to further understand this process.

Approximately half the women calling for appointments were not eligible to participate in our study because of the presence of 1 or more complicating factor. The most common reason was the presence of back pain, a complaint that commonly accompanies uncomplicated as well as complicated UTIs. Although the prevalence of acute pyelonephritis is very low, our protocol conservatively put women with this isolated complaint in a potentially high-risk group that required an office visit. It is quite likely that using a constellation of symptoms (such as back or flank pain plus fever or chills or nausea and vomiting, and so forth) would have allowed more women to be eligible. The participation rate was high among eligible women, improving the generalizability of the data. Although we enrolled predominantly white women (reflecting the ethnic mix of the participating practices), we believe the biologic responses in our study are not race dependent. We are not confident, however, that patient satisfaction data will extrapolate to other groups, since women in groups that have traditionally been underserved by the health care system may see telephone management as a way to shortchange them.

Our study was planned to have 80% power to detect important differences in the primary outcome variables but lacks sufficient sample size to determine if patients in either group were more likely to experience pyelonephritis or other complications. Since the specific therapy was similar in each group, one would suspect a similar rate of complications.

Twice as many patients in the telephone group were still symptomatic after 10 days, compared with those seen in the office. The small numbers in our study raise the possibility that clinically meaningful differences did not reach statistical significance. However, a closer look at the UTI scores suggests that only 1 of the 12 patients in the telephone group who reported persistent symptoms had a high score compared with 3 of the 6 control patients. This suggests that the severity of the persistent symptoms was quite low. Also, it raises the possibility that symptoms such as low back pain that were not captured by the UTI score or possibly not related to UTI were unimproved. We also believe that many of these women may have had other conditions causing their persistent symptoms. Finally, it is possible that these findings reflect a significant degree of statistical “noise” due to the wide confidence intervals associated with small studies. This is an area for further study.

This study used 7 days of antibiotic therapy. Currently, 3 days of therapy are increasingly used. Interestingly, more than half the women were still symptomatic on day 3. At the conclusion of the study, though, 75% of all women reported resolution of their symptoms. Although this discussion places this observation in a different context, it may raise potential concerns about whether 3-day therapy (while effective in delivering laboratory “cures”) may not provide enough relief to patients to be worth the tradeoffs.

Conclusions

This study demonstrates that managing uncomplicated UTIs in otherwise healthy women over the telephone has comparable outcomes and patient satisfaction with managing these women with an office visit. Whether symptom resolution is the same is not adequately answered by our study. More research on the optimal use of triage protocols for common acute conditions is needed in the primary care setting.

Acknowledgments

Our research was funded by the Blue Cross Blue Shield of Michigan Foundation grant # 231-II: a randomized clinical trial comparing telephone and usual care strategies for the management of suspected UTI in otherwise healthy adult women. We thank the following practices that participated in our study: Michigan State University Department of Family Practice, East Lansing; Order of St. Francis Medical Group, Escanaba; and Doctors Park Family Physicians, Escanaba. We are especially appreciative of the efforts of the office nurses and physician’s assistants who recruited and provided the telephone follow-up of the patients: Barb Bedient, LPN; Lisa Sweet, LPN; Debi Besson, RN; Grace Borkadi, PAC; Gloria Johnson, LPN; and Mary Baron, RN. Conflict of interest statement: Dr Ebell is editor of The Journal of Family Practice, Dr Hickner is an associate editor, and Dr Barry is an assistant editor. Therefore, the peer review process, including selection of reviewers, editorial review, editing, and the decision to accept or reject the manuscript was performed by Dr Bernard Ewigman, MD, MSPH, Associate Editor of JFP.

Millions of women with acute dysuria show up at offices, urgent care centers, and emergency departments for suspected urinary tract infections (UTIs), accounting for more than $1 billion in direct costs.¹ Since most UTIs are uncomplicated, numerous strategies have been proposed for managing them in more efficient and less costly ways. Berg² found 82 separate management strategies among 137 family physicians, with costs ranging from negligible to $250.

In a previous study³ we used a cost-utility analysis to demonstrate that among office-based approaches, empiric therapy of suspected UTIs was most cost-effective. This was robust over a wide range of assumptions, including prevalence, test characteristics, costs, complication rates, and outcomes. These findings were recently confirmed by Fenwick and colleagues⁴ in a British analysis. Among the approaches commonly used, telephone management has the potential for reducing cost, increasing convenience for patients, and reducing barriers to care. Saint and coworkers⁵ demonstrated that a clinical practice guideline using telephone-based management of presumed UTIs reduced the use of urine tests and increased the use of guideline-specified antibiotics. Although telephone management is common, we were unable to find any studies directly comparing that approach with office-based care.

We report on the first trial in which women with suspected UTIs were randomly assigned to telephone management or office-based management. The purpose of our study was to identify the prevalence of UTIs in women presenting with suspected UTIs, to determine if telephone management was comparable in outcomes with those of office-based care, and to determine if women whose infections were managed by telephone were satisfied with their care.

Methods

Setting

We enrolled patients at 6 primary care offices (including a walk-in clinic) of the Upper Peninsula Research Network (UPRNet) and the Department of Family Practice at Michigan State University in East Lansing. UPRNet is a primary care research network in the Upper Peninsula of Michigan.

Subjects

Nonpregnant women 18 years or older completed an interview to confirm eligibility when they called their primary care physicians with a suspected uncomplicated UTI. We defined a suspected uncomplicated UTI as a complaint of dysuria, pain on urination, complaint of urinary urgency and frequency, or the patient’s saying, “I think I have a bladder infection.” Patients with symptoms compatible with pyelonephritis (fever, chills, sweats, back or flank pain, or vomiting), vaginitis, or cervicitis (presence of a new or changed vaginal discharge) were excluded from our study. We also excluded women with diabetes, a previous history of pyelonephritis or other complicated UTI, a UTI in the preceding month, symptoms lasting longer than 14 days, and known kidney disease, anatomic abnormalities, or previous renal surgery. In addition, we excluded women receiving chemotherapy and those who had received antibiotics in the preceding month. Informed consent was obtained. Enrollment occurred between October of 1997 and March of 1999. All enrolled patients received $25 for participating in the study. The Michigan State University Committee on Research Involving Human Subjects approved our study.

Procedures and Measures

We stratified each practice, and using a central computer-generated random number, we blindly allocated patients to either treatment by telephone (telephone group) or office-based care (control group) by using an opaque envelope containing the group assignment. The envelope also provided instructions appropriate for the assigned group. We asked the telephone strategy patients to come to the clinic to leave a urine sample and pick up a prescription for an antibiotic; the control (usual care) patients were given a same-day appointment for a regular clinic visit.

We asked patients enrolled in both groups to rate the severity of urinary dysuria, urgency, frequency, function, and how they generally felt about their symptoms. Each was rated on a 10-point scale (10 was most severe). The telephone management patients were given a prescription for sulfamethoxazole/trimethoprim (800 mg/160 mg) twice daily for 7 days. If the patient was allergic to sulfa, she received a prescription for nitrofurantoin 100 mg twice daily for 7 days. Patients were required to have a urinalysis and urine culture before receiving the prescription. We asked the health care providers of the control patients to use their usual management strategies. The control group patients were also required to have a urinalysis and urine culture.

A nurse telephoned all women in both groups for follow-up 3 and 10 days after the time of enrollment. During these telephone calls we assessed patient status (including symptom scores and patient satisfaction) and whether the patient sought care with any practitioner since the initial time of enrollment. If UTI symptoms were still evident at either the 3- or 10-day follow-up, the patient was asked to make an appointment to see her practitioner. We recorded start and stop times for all interviews to assist in estimating and comparing costs of care for the 2 groups.

Primary Outcomes

The primary outcomes for our study were the UTI score and overall evaluation rating (OER)⁶ of the treatment experience. The UTI score is the sum of the ratings of severity of dysuria, urgency, frequency, function, and general symptoms (range=0-50). Since the UTI score has not been previously used, we pilot tested it on 20 women. These same women were reevaluated 2 to 3 days later. Test-retest reliability (measured by the overall correlation between the same items asked on 2 separate occasions with the order of questions changed) was 0.98. Validity (measured by the correlation coefficient between specific questions and a global question for the episode) was 0.52 for burning, 0.89 for frequency, 0.95 for urgency, 0.86 for interference with activities of daily living, and -0.76 for the actual interval between urination (smaller interval associated with greater overall episode severity).

The OER consisted of 2 questions about the overall quality of care and the outcome of care. Each question—answered as poor, fair, good, very good, or excellent—was scored from 1 to 5, respectively. The OER, a validated score,⁶ is a simple sum of the scores for these items (range 2-10, not normally distributed). Also, we asked the women about their overall satisfaction on a 5-point Likert scale.

Secondary Outcomes

We also evaluated urine culture results. We defined a negative culture as one with either no growth or less than 1000 colony-forming units. A positive culture had any growth of a single organism. A contaminated specimen (mixed flora) was rated as a negative culture, since this is how these are usually handled clinically.

Statistical Analysis

The data were entered into a database, and all entries were double-checked by one of the investigators for transcription errors. We analyzed the data using SAS software (version 7, SAS Institute, Cary, NC). Continuous variables (age, time variables, UTI score) were compared by treatment group using unpaired Student t tests. We used the Shapiro-Wilk test and the Kolmogorov-Smirnov test to confirm that the UTI score and changes in that score between assessments were normally distributed. We compared categorical variables (resolution of symptoms, culture results) by treatment group using a chi-square. We used Wilcoxon rank sum tests to compare ordinal data (OER) between the treatment groups.

For all sample size estimations we wanted to achieve 80% power with two-sided a of 0.05. To detect a 5-point difference in the UTI score, with an estimated a priori standard deviation (SD) of 5, we calculated that 16 subjects in each group are needed (21 for 90% power). With an SD of 7.5, 36 subjects in each group are needed. For patient satisfaction (OER), we estimated a priori that we would need 15 subjects in each group (19 for 90% power) to detect a 1-point difference with an estimated mean of 3.9 and a SD of 0.95.

Results

We identified 201 women with suspected UTIs of whom 99 were not eligible. Of the 102 eligible women, 30 declined to participate. We randomized 36 women to office care (control group) and 36 to telephone management. The subjects were predominantly white (n=62, 86%) and young (mean age=36.6 years, SD=12.3). Five patients had no culture results. Of the 67 remaining cultures, 4 (6%) were contaminated specimens, 20 (29.8%) had negative cultures, and 43 (64.2%) had positive cultures. Of the positive cultures, 34 (79.1%) grew Escherichia coli. Twenty-three of 34 (67.6%) cultures in the control group were positive compared with 20 of 33 (60.6%) in the telephone group (chi-square=0.3611; P=.55).

The groups were similar at baseline Table 1. On day 3 and day 10 there were no significant differences in the change in symptom scores or overall UTI score from baseline. We also found no difference in the change in urinary intervals from baseline. Table 2 shows these data. There was no significant difference in the overall evaluation rating. We also found no difference in satisfaction with care (median response was “very good” in the control group and “excellent” in the telephone group). These are shown in Table 3

On the third day after therapy was inititated, 20 of 33 (60.6%) of the control subjects had persistent urinary symptoms compared with 19 of 34 (55.8%, chi-square=0.1536; P=.70) in the telephone group. By day 10, 6 of 35 control patients (17.1%) had persistent symptoms, compared with 12 of 35 (34.3%) in the telephone group (chi-square=2.6923; P=0.1). Among the patients still symptomatic on the third day, culture results were available for 35, 11 (31%) of which were negative. Among those still symptomatic on the 10th day, 18 had available cultures, 9 (50%) of which were negative.

To evaluate the patients with persistent symptoms at the conclusion of the study we looked at the baseline and final UTI scores and baseline culture results. Six patients in the control group reported persistent symptoms. Three of these patients had final UTI scores less than 10, and 3 had UTI scores greater than 20. Two of these patients also had negative cultures. Twelve patients in the telephone group reported persistent symptoms; all but 1 had final UTI scores less than 10, and only 1 had a final UTI score greater than 20. Seven of the 12 patients had negative cultures at baseline.

We attempted to determine how office care differed from telephone care. Three patients in the control group received no antibiotics. Two of these had negative cultures, and no culture result was available for the third patient. All patients in the telephone group were prescribed antibiotics. Five control group patients who ultimately had positive cultures took antibiotics for less than 7 days, compared with only 3 in the telephone group. Among those receiving antibiotics, 30 of the control group patients received either sulfamethoxazole/trimethoprim or nitrofurantoin, and 3 received second-line agents. Because of allergies, 1 patient in the telephone group did not receive the planned therapy and received cephalexin instead.

We also evaluated the nursing time to administer various elements of the protocol. It took 2.5 minutes (SD=1.3) to determine eligibility to participate in the study and 5.3 minutes (SD=2.1) to enroll the subjects into the study. The nursing time for the day 3 follow-up took on average 5.6 minutes (SD=2.9) and 5.2 minutes (SD=2.0) on day 10.

Discussion

Although managing uncomplicated UTIs by telephone is a common practice in ambulatory primary care settings, we had no previous empiric evidence of its effectiveness compared with seeing patients in the office. In this randomized trial of office management versus management by telephone, two thirds of the women enrolled had culture-confirmed UTIs. The rate was similar in each group and mirrors that reported in the literature.⁷ We found no difference in improvement in symptom scores from baseline and no significant difference in overall satisfaction with the care provided or the outcome.

Gallagher and colleagues⁸ reported that when acute medical problems are triaged by nurses, patients are generally satisfied with care. However, UTIs represented only 5% of the telephone encounters. Delichatsios and coworkers⁹ similarly reported that patients calling to speak with the physician were generally satisfied with the advice given on the telephone, but they did not report outcomes related to specific conditions or therapies. Although 2 independent economic evaluations^3,4 have found empiric therapy to be cost-effective, neither included a strategy that avoids an office visit.

The direct cost of telephone management of uncomplicated UTIs is relatively low. It took only 2.5 minutes of nurse time to identify symptomatic women with risk factors for complicated UTIs who were good candidates for telephone management. This may cause a dilemma. Physicians practicing in predominantly fee-for-service settings will lose income by managing UTIs by telephone. In managed care settings, the financial incentives to reduce utilization make this practice inexpensive while simultaneously maintaining high patient satisfaction. Many physicians, however, complain about the complexity of the patients they now see, and having an occasional uncomplicated UTI might provide some breathing space on hectic days.

Limitations

We did not ask the practitioners who provided office-based care to alter their usual approach. By patient report, only 3 control patients received no antibiotics. This may reflect a knee-jerk response in which antibiotics are prescribed for all women with a suspected UTI. It may also reflect a very appropriate therapeutic threshold where physicians have a gestalt about the probability of a UTI that exceeds any diagnostic uncertainty. Although this has been described explicitly,¹⁰ we believe that seasoned clinicians do this implicitly. We did not attempt to open the “black box” to further understand this process.

Approximately half the women calling for appointments were not eligible to participate in our study because of the presence of 1 or more complicating factor. The most common reason was the presence of back pain, a complaint that commonly accompanies uncomplicated as well as complicated UTIs. Although the prevalence of acute pyelonephritis is very low, our protocol conservatively put women with this isolated complaint in a potentially high-risk group that required an office visit. It is quite likely that using a constellation of symptoms (such as back or flank pain plus fever or chills or nausea and vomiting, and so forth) would have allowed more women to be eligible. The participation rate was high among eligible women, improving the generalizability of the data. Although we enrolled predominantly white women (reflecting the ethnic mix of the participating practices), we believe the biologic responses in our study are not race dependent. We are not confident, however, that patient satisfaction data will extrapolate to other groups, since women in groups that have traditionally been underserved by the health care system may see telephone management as a way to shortchange them.

Our study was planned to have 80% power to detect important differences in the primary outcome variables but lacks sufficient sample size to determine if patients in either group were more likely to experience pyelonephritis or other complications. Since the specific therapy was similar in each group, one would suspect a similar rate of complications.

Twice as many patients in the telephone group were still symptomatic after 10 days, compared with those seen in the office. The small numbers in our study raise the possibility that clinically meaningful differences did not reach statistical significance. However, a closer look at the UTI scores suggests that only 1 of the 12 patients in the telephone group who reported persistent symptoms had a high score compared with 3 of the 6 control patients. This suggests that the severity of the persistent symptoms was quite low. Also, it raises the possibility that symptoms such as low back pain that were not captured by the UTI score or possibly not related to UTI were unimproved. We also believe that many of these women may have had other conditions causing their persistent symptoms. Finally, it is possible that these findings reflect a significant degree of statistical “noise” due to the wide confidence intervals associated with small studies. This is an area for further study.

This study used 7 days of antibiotic therapy. Currently, 3 days of therapy are increasingly used. Interestingly, more than half the women were still symptomatic on day 3. At the conclusion of the study, though, 75% of all women reported resolution of their symptoms. Although this discussion places this observation in a different context, it may raise potential concerns about whether 3-day therapy (while effective in delivering laboratory “cures”) may not provide enough relief to patients to be worth the tradeoffs.

Conclusions

This study demonstrates that managing uncomplicated UTIs in otherwise healthy women over the telephone has comparable outcomes and patient satisfaction with managing these women with an office visit. Whether symptom resolution is the same is not adequately answered by our study. More research on the optimal use of triage protocols for common acute conditions is needed in the primary care setting.

Acknowledgments

Our research was funded by the Blue Cross Blue Shield of Michigan Foundation grant # 231-II: a randomized clinical trial comparing telephone and usual care strategies for the management of suspected UTI in otherwise healthy adult women. We thank the following practices that participated in our study: Michigan State University Department of Family Practice, East Lansing; Order of St. Francis Medical Group, Escanaba; and Doctors Park Family Physicians, Escanaba. We are especially appreciative of the efforts of the office nurses and physician’s assistants who recruited and provided the telephone follow-up of the patients: Barb Bedient, LPN; Lisa Sweet, LPN; Debi Besson, RN; Grace Borkadi, PAC; Gloria Johnson, LPN; and Mary Baron, RN. Conflict of interest statement: Dr Ebell is editor of The Journal of Family Practice, Dr Hickner is an associate editor, and Dr Barry is an assistant editor. Therefore, the peer review process, including selection of reviewers, editorial review, editing, and the decision to accept or reject the manuscript was performed by Dr Bernard Ewigman, MD, MSPH, Associate Editor of JFP.

Millions of women with acute dysuria show up at offices, urgent care centers, and emergency departments for suspected urinary tract infections (UTIs), accounting for more than $1 billion in direct costs.¹ Since most UTIs are uncomplicated, numerous strategies have been proposed for managing them in more efficient and less costly ways. Berg² found 82 separate management strategies among 137 family physicians, with costs ranging from negligible to $250.

In a previous study³ we used a cost-utility analysis to demonstrate that among office-based approaches, empiric therapy of suspected UTIs was most cost-effective. This was robust over a wide range of assumptions, including prevalence, test characteristics, costs, complication rates, and outcomes. These findings were recently confirmed by Fenwick and colleagues⁴ in a British analysis. Among the approaches commonly used, telephone management has the potential for reducing cost, increasing convenience for patients, and reducing barriers to care. Saint and coworkers⁵ demonstrated that a clinical practice guideline using telephone-based management of presumed UTIs reduced the use of urine tests and increased the use of guideline-specified antibiotics. Although telephone management is common, we were unable to find any studies directly comparing that approach with office-based care.

We report on the first trial in which women with suspected UTIs were randomly assigned to telephone management or office-based management. The purpose of our study was to identify the prevalence of UTIs in women presenting with suspected UTIs, to determine if telephone management was comparable in outcomes with those of office-based care, and to determine if women whose infections were managed by telephone were satisfied with their care.

Methods

Setting

We enrolled patients at 6 primary care offices (including a walk-in clinic) of the Upper Peninsula Research Network (UPRNet) and the Department of Family Practice at Michigan State University in East Lansing. UPRNet is a primary care research network in the Upper Peninsula of Michigan.

Subjects

Nonpregnant women 18 years or older completed an interview to confirm eligibility when they called their primary care physicians with a suspected uncomplicated UTI. We defined a suspected uncomplicated UTI as a complaint of dysuria, pain on urination, complaint of urinary urgency and frequency, or the patient’s saying, “I think I have a bladder infection.” Patients with symptoms compatible with pyelonephritis (fever, chills, sweats, back or flank pain, or vomiting), vaginitis, or cervicitis (presence of a new or changed vaginal discharge) were excluded from our study. We also excluded women with diabetes, a previous history of pyelonephritis or other complicated UTI, a UTI in the preceding month, symptoms lasting longer than 14 days, and known kidney disease, anatomic abnormalities, or previous renal surgery. In addition, we excluded women receiving chemotherapy and those who had received antibiotics in the preceding month. Informed consent was obtained. Enrollment occurred between October of 1997 and March of 1999. All enrolled patients received $25 for participating in the study. The Michigan State University Committee on Research Involving Human Subjects approved our study.

Procedures and Measures

We stratified each practice, and using a central computer-generated random number, we blindly allocated patients to either treatment by telephone (telephone group) or office-based care (control group) by using an opaque envelope containing the group assignment. The envelope also provided instructions appropriate for the assigned group. We asked the telephone strategy patients to come to the clinic to leave a urine sample and pick up a prescription for an antibiotic; the control (usual care) patients were given a same-day appointment for a regular clinic visit.

We asked patients enrolled in both groups to rate the severity of urinary dysuria, urgency, frequency, function, and how they generally felt about their symptoms. Each was rated on a 10-point scale (10 was most severe). The telephone management patients were given a prescription for sulfamethoxazole/trimethoprim (800 mg/160 mg) twice daily for 7 days. If the patient was allergic to sulfa, she received a prescription for nitrofurantoin 100 mg twice daily for 7 days. Patients were required to have a urinalysis and urine culture before receiving the prescription. We asked the health care providers of the control patients to use their usual management strategies. The control group patients were also required to have a urinalysis and urine culture.

A nurse telephoned all women in both groups for follow-up 3 and 10 days after the time of enrollment. During these telephone calls we assessed patient status (including symptom scores and patient satisfaction) and whether the patient sought care with any practitioner since the initial time of enrollment. If UTI symptoms were still evident at either the 3- or 10-day follow-up, the patient was asked to make an appointment to see her practitioner. We recorded start and stop times for all interviews to assist in estimating and comparing costs of care for the 2 groups.

Primary Outcomes

The primary outcomes for our study were the UTI score and overall evaluation rating (OER)⁶ of the treatment experience. The UTI score is the sum of the ratings of severity of dysuria, urgency, frequency, function, and general symptoms (range=0-50). Since the UTI score has not been previously used, we pilot tested it on 20 women. These same women were reevaluated 2 to 3 days later. Test-retest reliability (measured by the overall correlation between the same items asked on 2 separate occasions with the order of questions changed) was 0.98. Validity (measured by the correlation coefficient between specific questions and a global question for the episode) was 0.52 for burning, 0.89 for frequency, 0.95 for urgency, 0.86 for interference with activities of daily living, and -0.76 for the actual interval between urination (smaller interval associated with greater overall episode severity).

The OER consisted of 2 questions about the overall quality of care and the outcome of care. Each question—answered as poor, fair, good, very good, or excellent—was scored from 1 to 5, respectively. The OER, a validated score,⁶ is a simple sum of the scores for these items (range 2-10, not normally distributed). Also, we asked the women about their overall satisfaction on a 5-point Likert scale.

Secondary Outcomes

We also evaluated urine culture results. We defined a negative culture as one with either no growth or less than 1000 colony-forming units. A positive culture had any growth of a single organism. A contaminated specimen (mixed flora) was rated as a negative culture, since this is how these are usually handled clinically.

Statistical Analysis

The data were entered into a database, and all entries were double-checked by one of the investigators for transcription errors. We analyzed the data using SAS software (version 7, SAS Institute, Cary, NC). Continuous variables (age, time variables, UTI score) were compared by treatment group using unpaired Student t tests. We used the Shapiro-Wilk test and the Kolmogorov-Smirnov test to confirm that the UTI score and changes in that score between assessments were normally distributed. We compared categorical variables (resolution of symptoms, culture results) by treatment group using a chi-square. We used Wilcoxon rank sum tests to compare ordinal data (OER) between the treatment groups.

For all sample size estimations we wanted to achieve 80% power with two-sided a of 0.05. To detect a 5-point difference in the UTI score, with an estimated a priori standard deviation (SD) of 5, we calculated that 16 subjects in each group are needed (21 for 90% power). With an SD of 7.5, 36 subjects in each group are needed. For patient satisfaction (OER), we estimated a priori that we would need 15 subjects in each group (19 for 90% power) to detect a 1-point difference with an estimated mean of 3.9 and a SD of 0.95.

Results

We identified 201 women with suspected UTIs of whom 99 were not eligible. Of the 102 eligible women, 30 declined to participate. We randomized 36 women to office care (control group) and 36 to telephone management. The subjects were predominantly white (n=62, 86%) and young (mean age=36.6 years, SD=12.3). Five patients had no culture results. Of the 67 remaining cultures, 4 (6%) were contaminated specimens, 20 (29.8%) had negative cultures, and 43 (64.2%) had positive cultures. Of the positive cultures, 34 (79.1%) grew Escherichia coli. Twenty-three of 34 (67.6%) cultures in the control group were positive compared with 20 of 33 (60.6%) in the telephone group (chi-square=0.3611; P=.55).

The groups were similar at baseline Table 1. On day 3 and day 10 there were no significant differences in the change in symptom scores or overall UTI score from baseline. We also found no difference in the change in urinary intervals from baseline. Table 2 shows these data. There was no significant difference in the overall evaluation rating. We also found no difference in satisfaction with care (median response was “very good” in the control group and “excellent” in the telephone group). These are shown in Table 3

On the third day after therapy was inititated, 20 of 33 (60.6%) of the control subjects had persistent urinary symptoms compared with 19 of 34 (55.8%, chi-square=0.1536; P=.70) in the telephone group. By day 10, 6 of 35 control patients (17.1%) had persistent symptoms, compared with 12 of 35 (34.3%) in the telephone group (chi-square=2.6923; P=0.1). Among the patients still symptomatic on the third day, culture results were available for 35, 11 (31%) of which were negative. Among those still symptomatic on the 10th day, 18 had available cultures, 9 (50%) of which were negative.

To evaluate the patients with persistent symptoms at the conclusion of the study we looked at the baseline and final UTI scores and baseline culture results. Six patients in the control group reported persistent symptoms. Three of these patients had final UTI scores less than 10, and 3 had UTI scores greater than 20. Two of these patients also had negative cultures. Twelve patients in the telephone group reported persistent symptoms; all but 1 had final UTI scores less than 10, and only 1 had a final UTI score greater than 20. Seven of the 12 patients had negative cultures at baseline.

We attempted to determine how office care differed from telephone care. Three patients in the control group received no antibiotics. Two of these had negative cultures, and no culture result was available for the third patient. All patients in the telephone group were prescribed antibiotics. Five control group patients who ultimately had positive cultures took antibiotics for less than 7 days, compared with only 3 in the telephone group. Among those receiving antibiotics, 30 of the control group patients received either sulfamethoxazole/trimethoprim or nitrofurantoin, and 3 received second-line agents. Because of allergies, 1 patient in the telephone group did not receive the planned therapy and received cephalexin instead.

We also evaluated the nursing time to administer various elements of the protocol. It took 2.5 minutes (SD=1.3) to determine eligibility to participate in the study and 5.3 minutes (SD=2.1) to enroll the subjects into the study. The nursing time for the day 3 follow-up took on average 5.6 minutes (SD=2.9) and 5.2 minutes (SD=2.0) on day 10.

Discussion

Although managing uncomplicated UTIs by telephone is a common practice in ambulatory primary care settings, we had no previous empiric evidence of its effectiveness compared with seeing patients in the office. In this randomized trial of office management versus management by telephone, two thirds of the women enrolled had culture-confirmed UTIs. The rate was similar in each group and mirrors that reported in the literature.⁷ We found no difference in improvement in symptom scores from baseline and no significant difference in overall satisfaction with the care provided or the outcome.

Gallagher and colleagues⁸ reported that when acute medical problems are triaged by nurses, patients are generally satisfied with care. However, UTIs represented only 5% of the telephone encounters. Delichatsios and coworkers⁹ similarly reported that patients calling to speak with the physician were generally satisfied with the advice given on the telephone, but they did not report outcomes related to specific conditions or therapies. Although 2 independent economic evaluations^3,4 have found empiric therapy to be cost-effective, neither included a strategy that avoids an office visit.

The direct cost of telephone management of uncomplicated UTIs is relatively low. It took only 2.5 minutes of nurse time to identify symptomatic women with risk factors for complicated UTIs who were good candidates for telephone management. This may cause a dilemma. Physicians practicing in predominantly fee-for-service settings will lose income by managing UTIs by telephone. In managed care settings, the financial incentives to reduce utilization make this practice inexpensive while simultaneously maintaining high patient satisfaction. Many physicians, however, complain about the complexity of the patients they now see, and having an occasional uncomplicated UTI might provide some breathing space on hectic days.

Limitations

We did not ask the practitioners who provided office-based care to alter their usual approach. By patient report, only 3 control patients received no antibiotics. This may reflect a knee-jerk response in which antibiotics are prescribed for all women with a suspected UTI. It may also reflect a very appropriate therapeutic threshold where physicians have a gestalt about the probability of a UTI that exceeds any diagnostic uncertainty. Although this has been described explicitly,¹⁰ we believe that seasoned clinicians do this implicitly. We did not attempt to open the “black box” to further understand this process.

Approximately half the women calling for appointments were not eligible to participate in our study because of the presence of 1 or more complicating factor. The most common reason was the presence of back pain, a complaint that commonly accompanies uncomplicated as well as complicated UTIs. Although the prevalence of acute pyelonephritis is very low, our protocol conservatively put women with this isolated complaint in a potentially high-risk group that required an office visit. It is quite likely that using a constellation of symptoms (such as back or flank pain plus fever or chills or nausea and vomiting, and so forth) would have allowed more women to be eligible. The participation rate was high among eligible women, improving the generalizability of the data. Although we enrolled predominantly white women (reflecting the ethnic mix of the participating practices), we believe the biologic responses in our study are not race dependent. We are not confident, however, that patient satisfaction data will extrapolate to other groups, since women in groups that have traditionally been underserved by the health care system may see telephone management as a way to shortchange them.

Our study was planned to have 80% power to detect important differences in the primary outcome variables but lacks sufficient sample size to determine if patients in either group were more likely to experience pyelonephritis or other complications. Since the specific therapy was similar in each group, one would suspect a similar rate of complications.

Twice as many patients in the telephone group were still symptomatic after 10 days, compared with those seen in the office. The small numbers in our study raise the possibility that clinically meaningful differences did not reach statistical significance. However, a closer look at the UTI scores suggests that only 1 of the 12 patients in the telephone group who reported persistent symptoms had a high score compared with 3 of the 6 control patients. This suggests that the severity of the persistent symptoms was quite low. Also, it raises the possibility that symptoms such as low back pain that were not captured by the UTI score or possibly not related to UTI were unimproved. We also believe that many of these women may have had other conditions causing their persistent symptoms. Finally, it is possible that these findings reflect a significant degree of statistical “noise” due to the wide confidence intervals associated with small studies. This is an area for further study.

This study used 7 days of antibiotic therapy. Currently, 3 days of therapy are increasingly used. Interestingly, more than half the women were still symptomatic on day 3. At the conclusion of the study, though, 75% of all women reported resolution of their symptoms. Although this discussion places this observation in a different context, it may raise potential concerns about whether 3-day therapy (while effective in delivering laboratory “cures”) may not provide enough relief to patients to be worth the tradeoffs.

Conclusions

This study demonstrates that managing uncomplicated UTIs in otherwise healthy women over the telephone has comparable outcomes and patient satisfaction with managing these women with an office visit. Whether symptom resolution is the same is not adequately answered by our study. More research on the optimal use of triage protocols for common acute conditions is needed in the primary care setting.

Acknowledgments

Our research was funded by the Blue Cross Blue Shield of Michigan Foundation grant # 231-II: a randomized clinical trial comparing telephone and usual care strategies for the management of suspected UTI in otherwise healthy adult women. We thank the following practices that participated in our study: Michigan State University Department of Family Practice, East Lansing; Order of St. Francis Medical Group, Escanaba; and Doctors Park Family Physicians, Escanaba. We are especially appreciative of the efforts of the office nurses and physician’s assistants who recruited and provided the telephone follow-up of the patients: Barb Bedient, LPN; Lisa Sweet, LPN; Debi Besson, RN; Grace Borkadi, PAC; Gloria Johnson, LPN; and Mary Baron, RN. Conflict of interest statement: Dr Ebell is editor of The Journal of Family Practice, Dr Hickner is an associate editor, and Dr Barry is an assistant editor. Therefore, the peer review process, including selection of reviewers, editorial review, editing, and the decision to accept or reject the manuscript was performed by Dr Bernard Ewigman, MD, MSPH, Associate Editor of JFP.

Busy clinicians have come to rely on the office laboratory as an important tool in providing timely, efficient, and high-quality patient care. However, concerns have been raised about poor accuracy in physician office laboratories and financial incentives that encourage overuse of tests.^1-7 Recent studies have documented inaccurate office laboratory testing.^8,9 Previously, the Commission on Office Laboratory Accreditation concluded that approximately 5% of office laboratories have serious deficiencies involving quality control, instrument maintenance, specimen management, and proficiency testing.¹⁰ Similar concerns led to the passage of the Clinical Laboratory Improvement Act of 1988 (CLIA-88), which was implemented in the fall of 1994.¹¹ Before CLIA-88, physician office laboratory quality regulations were decided at the state level.¹² Both state and federal regulations require adherence to specific measures (eg, proficiency testing, quality control, and quality assurance) to operate an office laboratory.

The critics of quality regulations assert that the measures might actually decrease the quality of care. They note that the regulations impose a fixed cost on the physician’s practice that must be spent regardless of how busy the laboratory is. The additional costs of compliance could discourage physicians from offering tests that would otherwise be done.^13,14 This reduction in test availability might occur in both high- and low-quality laboratories. Therefore, regulations aimed at improving quality might have the unintended effect of reducing the availability and use of clinically beneficial laboratory tests, thereby decreasing the overall potential benefit of the regulations.

We previously surveyed primary care physicians about their clinical approach to patients with a possible uncomplicated urinary tract infection (UTI).¹⁵ That 1994-1995 survey included primary care physicians in 4 states and contained detailed information on their work settings and clinical behavior concerning cases of UTI. In this paper we use those survey results to examine the potential effects of office laboratory quality regulations and test availability on testing for UTIs. This is possible because one of the states surveyed, Pennsylvania, has had statewide physician office laboratory quality regulations similar to CLIA-88 in place since the mid-1970s.^12,16 All office laboratories in Pennsylvania were required to register with the state, have a written procedure manual, perform quality and proficiency testing, and comply with state regulations, regardless of the level of testing.¹⁶ Office laboratories in the other 3 states were not regulated until the implementation of CLIA-88 in late 1994. Thus, the survey includes 1 regulated state and 3 states that were just beginning to be regulated. This should provide a conservative estimate of the differences between a regulated and a nonregulated situation, since some of the physicians in the previously unregulated states may already have adjusted to CLIA-88 when our survey was done. The adjustment to CLIA-88 did not occur instantaneously; laboratories have continued to move toward a higher percentage of waived and provider-performed microscopy laboratories (from 57.8% of all laboratories in 1995 to 75% in July 2000) since CLIA-88 was implemented.¹⁷ We sought to determine if office laboratory testing capabilities or availability differs by state and if self-reported use of the tests in diagnosing an uncomplicated UTI is related to the availability of the tests.

Methods

We examined the results of an 88-item survey concerning each physician’s clinical approach to patients with a possible UTI. We mailed the survey in 1994 and 1995 to practicing primary care physicians in 4 specialties that commonly treat adults with uncomplicated UTIs: general internal medicine, family practice, obstetrics and gynecology, and emergency medicine. The emergency medicine physicians were excluded from this analysis, since most of these physicians are hospital based and do not have an office laboratory in the same sense as the other specialties. To obtain a geographically diverse sample, we surveyed physicians from 4 states: Alabama, Nebraska, Minnesota, and Pennsylvania. The names and addresses of potential respondents were obtained from the physician masterfile of the American Medical Association (AMA). Since the AMA masterfile has limited data on physician characteristics (eg, the physician’s specialty), all data in our analysis were based on self-reported survey results. We surveyed the entire population of eligible physicians in each state except Pennsylvania; a random sample was chosen from Pennsylvania because of the large number of practicing physicians in that state.

The survey asked detailed questions about each physician’s clinical approach to a 30-year-old woman with dysuria, a presentation suggestive but not diagnostic of a UTI.^18,19 The main outcomes included whether specific tests (urine dipstick, microscopic urinalysis [UA], wet prep, and urine culture) were available in the office and whether these tests were used when diagnosing a UTI in the given patient. In both parts of the analysis we controlled for possible confounding variables. These included items reflecting the physician’s belief in the usefulness of clinical and laboratory information in diagnosing UTIs and physician and practice characteristics. The variables used in the analyses are shown in Table 1.

In the first part of the analysis we determined whether practicing in a regulated state (Pennsylvania) is associated with changes in the likelihood of having tests available in the office. The 4 outcome variables in this part of the analysis included the presence or absence of the dipstick, microscopic UA, wet prep, and urine culture. Thus, test availability was analyzed using binary dependent variables (yes=the physician reported the test in the office; no=otherwise). We did multivariate logistic regressions because of the binary nature of the dependent variables. Variables explaining the presence or absence of the test in the office included the key variable of interest: the physician’s state of residence (either in Pennsylvania or 1 of the previously unregulated states) and the group of variables reflecting the clinical beliefs concerning history, physical examination, and test usefulness in diagnosing UTI in our hypothetical patient, as well as physician, practice, and community characteristics. The explanatory variables were included in the regression models to control for any factors that might confound the effects of being from Pennsylvania. We hypothesized that tests would be found less frequently in Pennsylvania when controlling for other factors.

In the second part of the analysis, we examined whether the availability of tests in the office is related to their use in diagnosing UTIs in the hypothetical patient. The outcome variables in the second part were the self-reported frequency of ordering microscopic UA, urine culture, and the urine dipstick test (the leukocyte esterase and/or the nitrite test). Test use was analyzed as a binary variable (yes=the physician sometimes or usually performed the test; no=the physician rarely did the test). We did not have information on the physician’s use of the wet prep test, so we could not analyze the relationship between availablity and use for the wet prep. Because we analyzed the variable representing test use as a binary variable, we also used logistic regression for the analysis in the second part. In each regression, the variables explaining the frequency of using the test included 3 outcome measures from the first part of the analysis (whether the dipstick, micro UA, or culture was available in the office), and the same control variables reflecting physician beliefs and the personal and practice characteristics used in the first part of the analysis. We hypothesized that test availability would be related to its use in diagnosing UTIs after controlling for other relevant factors.

We used the Stata statistical software program²⁰ for the analysis.

Results

A total of 8942 surveys were sent out. There were 2172 usable surveys returned. After excluding the responses from the 274 emergency medicine physicians, we analyzed the remaining 1898 responses. We were able to compare respondents and nonrespondents on 3 demographic characteristics: sex, board-certification status, and length of time since graduating from medical school. The survey responders were more likely to be board certified but otherwise were similar to nonrespondents Table 2. Comparison of response patterns in Pennsylvania and the other 3 states using logistic regression indicates that Pennsylvania respondents were less likely to be men than respondents from other states. The magnitude of these differences, however, appears small.

Analysis of Test Availability

The reported prevalence of the various diagnostic tests in office laboratories is shown in Table 3. The dipstick is the most common test; more than 90% of respondents reported having it in their offices. The culture was the least available test, and the prevalence of both the microscopic UA and wet prep were intermediate between the culture and the dipstick. Simple unadjusted comparisons indicate that for each test physicians’ offices in Pennsylvania were less likely to report that they have the test available than were physicians in the unregulated states (P <.0001).

Table 4 shows the determinants of test availability from the multivariate logistic regressions. After controlling for possibly confounding factors, physicians from Pennsylvania were much less likely to have each of the tests in the office. The odds ratios (ORs) range from 0.20 (for the microscopic UA) to 0.35 (for the dipstick). Each of these results is statistically significant.

A number of control variables were significantly related to test availability.* Important findings include the effects of the physician’s specialty, clinical beliefs, and the estimated number of patients with dysuria seen per week. General internal medicine physicians were less likely than family practitioners to have each of the tests in the office (OR = 0.14, 0.60, 0.35, 0.57 for the dipstick, micro UA, wet prep, and culture; all P values <.05), while obstetrician-gynecologists were significantly less likely than family practitioners to have a dipstick, micro UA, and culture (OR = 0.14, 0.24, 0.44; all P values <.05) but more likely to have a wet prep available (OR=5.46; P=.0001). Increased belief in the importance of the leukocyte eserase test for diagnosing UTIs was associated with an increase in dipstick availability (OR=3.40; P=.0001), and increased belief in the importance of the microscopic white cell and red cell readings for diagnosing UTIs was associated with an increase in micro UA availability (OR = 1.95, 1.83; P = .01, .001, respectively). Seeing an additional patient per week with dysuria was associated with an increased availability for the dipstick (OR=1.12; P=.04), micro UA (OR=1.06; P=.01), and wet prep (OR=1.06; P=.02).

The overall explanatory power of these models was fairly good. The C-statistic, representing the area under the receiver operating characteristic curve, ranged from 0.75 for the urine culture to 0.93 for the dipstick (a C-statistic of 0.5 indicates that the model is no better than random chance at predicting the outcome, while a C-statistic of 1 indicates perfect discriminating ability).

Test Availability and the Diagnostic Approach

The relationship between test availability and use demonstrates that the availability of some tests is associated with an increase in their use, while sometimes the availability of tests is associated with an increase or decrease in the use of other tests Table 5. That is, tests will sometimes substitute for or complement other tests when diagnosing UTIs. Physicians with a microscopic UA in the office were more likely to report ordering a microscopic UA. The presence of the urine culture increased microscopic UA use, but the dipstick, when available, appeared to substitute for the microscopic UA.

The reported frequency of ordering urine cultures was increased by the availability of a culture but was decreased when respondents reported having a microscopic UA or dipstick available.

The dipstick was the only test for which use was not significantly related to the presence of any of the diagnostic tests; the availability of the dipstick appears to increase its use, but this relationship was not significant.

Even though test availability is often related to use, the physician may not be the person actually doing the tests. Of the 1492 physicians who report sometimes or usually ordering microscopic UAs, 363 do the procedure themselves, and 803 use ancillary personnel in their office. The remainder (326) have the procedure done outside the office. In contrast, only 25 of the physicians who report sometimes or usually ordering a urine culture do the test themselves. Three hundred forty-two use ancillary help in the office, and 243 have the culture done at an outside laboratory.

Several physician and practice characteristics were significantly related to the frequency of ordering each of the diagnostic tests. Being an obstetrician-gynecologist relative to family practitioner decreased microscopic UA use (OR=0.41; P=.0001), while a greater belief in the value of the leukocyte esterase test decreased microscopic UA use (OR=0.70; P=.01), and greater belief in the microscopic white cell reading increased use of the microscopic UA (OR=3.12; P=.0001). Increased use of urine cultures was associated with being in a city with a population greater than 100,000 relative to a town of less than 10,000 (OR=1.54; P=.02) and being a government employee relative to being in private practice (OR=9.31; P=.0001), while culture use was decreased in practice sizes of more than 25 physicians relative to being in solo practice (OR=0.39; P=.004). The dipstick was used more commonly when physicians had a greater belief in the value of the leukocyte esterase (OR=5.15; P=.0001) and nitrite (OR=2.68; P=.0001) results. Interestingly, the physician’s specialty was generally not significantly related to test use (apart from the difference between obstetrician-gynecologists and family practitioners in micro UA use), in contrast to the consistent specialty differences found for test availability.

The overall power of each of these models to explain the frequency of test use was moderately good. The C-statistics ranged from 0.69 for ordering a urine culture to 0.82 for ordering the dipstick.

Discussion

This study has 2 main findings. First, for each of the office laboratory tests examined, we found significant differences in test availability between unregulated states and Pennsylvania, a state with a long-standing physician office laboratory regulatory system. The overall pattern is one of diminished test availability in Pennsylvania, after controlling for other explanatory influences. This finding was present for every test we examined, regardless of the cost of performing the test or the level at which it was regulated. This supports the hypothesis that tests are available less often in a regulated state.

The second main finding is that the physician’s diagnostic approach to UTIs was related to the in-office availability of tests. This finding supports our second hypothesis. However, the relationship between test and availability and use was more complex than suspected. Not only was the self-reported use of some tests (microscopic UA and culture) related to their availability, sometimes the availability of a test was associated with a change in the self-reported frequency of using other tests. Thus, some tests appear to substitute for others (availability of the dipstick reduces use of microscopic UAs and cultures, and availability of the microscopic UA decreases culture ordering), while the availability of the culture was associated with increased use of the microscopic UA. We also found significant relationships between the physician’s belief in the usefulness of test results in diagnosing UTI with both the availability and subsequent use of the diagnostic tests.

The multivariate analyses indicated that physician and practice characteristics were related to the availability and use of diagnostic tests. This finding is consistent with previous studies reporting numerous factors influencing the use of laboratory tests.^21-23 To our knowledge, the association between physician belief in the value of the test in diagnosing a condition and test availability and use has not been documented previously. Thus, the influence of quality regulations on test prevalence and use will be modulated by the configuration of these other factors.

This study differs from previous studies on laboratory use in that we have been able to control for many relevant explanatory variables, including the physician’s belief in the diagnostic importance of elements from the patient’s history, physical examination, and test results. Many of these items add significant predictive power in a “logical” direction. For example, the availability of many tests was related to the physician’s belief in the value of the results provided by that test, and these beliefs were related to the physician’s diagnostic approach. Although it is not possible to tell from the survey if the physician’s belief in the usefulness of the test is causally related to the availability of a test, the results indicate the need to control for physician-specific opinions and beliefs when examining laboratory test use.

It is not possible to determine the net effect of quality regulations on patient welfare from this study. In theory, a reduction in test availability and subsequent use of a less accurate diagnostic approach is offset by higher quality and accuracy in those offices still offering tests.²⁴ We did not measure the quality or accuracy of the physician’s laboratory testing. To the extent that poor-quality laboratories are closed, regulations should improve overall welfare. Alternatively, if the reduction in test availability occurs in both good- and bad-quality laboratories, then the net effects on patient outcomes may be negative. A recent cost-effectiveness study of tests used to diagnose UTIs²⁵ indicated that the microscopic UA was reasonably cost-effective ($2964 per quality-adjusted life month), but the low sensitivity of the dipstick resulted in a poor cost-effectiveness ratio ($48,460 per quality-adjusted life month). Thus, to the extent that the highly prevalent dipstick is substituting for more cost-effective tests, as our results suggest, regulations might lead to less cost-effective care.

Our analysis suggests that changing the availability of tests may change physicians’ diagnostic approaches in ways that are difficult to predict a priori. For example, reducing the availability of the microscopic UA might lead to greater use of the more expensive urine culture. When many tests are reduced in availability, it is difficult to anticipate how the care of patients will change, since tests will substitute for or complement the use of other tests in ways that are difficult to predict.

Limitations

This study has several limitations. First, it is based on survey data that may not reflect actual behavior. However, the results are plausible, internally consistent, and consistent with previous research findings documenting substantial variation in practice patterns in managing UTIs.²⁶ Our focus on how physicians treat an uncomplicated basic UTI limits the possibility of eliciting responses based on unmeasured patient characteristics. Second, the overall response rate was low, which might raise concerns that the respondents represent an atypical or biased group of physicians. A comparison of respondents and nonrespondents along the 3 characteristics we had access to did not reveal any striking differences. And, when compared with nationwide physician work force statistics collected by the AMA in 1994, the physicians in this study had similar hours per week spent in patient care (46.2 in our survey vs 48.2 in the AMA survey) and estimated level of capitation (19.3% in our survey vs 16.7% in the AMA survey).²⁷ The estimated volume of patients with UTI was also similar to previous surveys of primary care physicians (approximately 6 per week).²⁶ Thus, the respondents in this survey do not seem to represent an atypical or unusual group of physicians. Other limitations include the fact that physicians responding to the survey might not be responsible for deciding whether to acquire or maintain tests in the office. Likewise, the survey does not contain information on other factors in the physician’s local market that might influence the supply or demand for office tests. These effects, however, would have to be quite large and disproportionately affect Pennsylvania respondents to confound our findings.

Finally, since the analysis is based on cross-sectional data, the results do not give any indication of how rapidly in-office testing declines after the implementation of quality regulations. However, as mentioned before, statistics indicate that the mix of tests in physicians’ offices has continued to move towards simpler tests since CLIA-88 was implemented. Our study provides a snapshot of the circumstances as CLIA-88 was first implemented; it would be worthwhile to assess changes in test prevalence and use in our survey population over the last several years. It may be the case that other changes underway in health care delivery (eg, improving technology, vertical and horizontal integration of physician groups, and growth of managed care) may hasten or reverse the changes in office laboratory availability implied in this study.

Conclusions

The availability of common office tests used in the diagnosis of UTI appears to be related to many factors, including the presence of office laboratory regulations. Lower availability of tests, in turn, was associated with the diagnostic approach when treating patients with a possible UTI. The net effects on patient welfare are difficult to predict, but our findings raise a serious concern about the possibility of a detrimental overall effect on patient welfare. Determining the full effect of quality regulations will require careful study to measure the changes in treatment patterns, treatment quality, and patient outcomes subsequent to the implementation of increased levels of regulation.

Acknowledgments

This research was funded by the Department of Internal Medicine at the University of Nebraska Medical Center, Omaha, and by the Department of Internal Medicine at Abington Memorial Hospital, Philadelphia, Pennsylvania.

Busy clinicians have come to rely on the office laboratory as an important tool in providing timely, efficient, and high-quality patient care. However, concerns have been raised about poor accuracy in physician office laboratories and financial incentives that encourage overuse of tests.^1-7 Recent studies have documented inaccurate office laboratory testing.^8,9 Previously, the Commission on Office Laboratory Accreditation concluded that approximately 5% of office laboratories have serious deficiencies involving quality control, instrument maintenance, specimen management, and proficiency testing.¹⁰ Similar concerns led to the passage of the Clinical Laboratory Improvement Act of 1988 (CLIA-88), which was implemented in the fall of 1994.¹¹ Before CLIA-88, physician office laboratory quality regulations were decided at the state level.¹² Both state and federal regulations require adherence to specific measures (eg, proficiency testing, quality control, and quality assurance) to operate an office laboratory.

The critics of quality regulations assert that the measures might actually decrease the quality of care. They note that the regulations impose a fixed cost on the physician’s practice that must be spent regardless of how busy the laboratory is. The additional costs of compliance could discourage physicians from offering tests that would otherwise be done.^13,14 This reduction in test availability might occur in both high- and low-quality laboratories. Therefore, regulations aimed at improving quality might have the unintended effect of reducing the availability and use of clinically beneficial laboratory tests, thereby decreasing the overall potential benefit of the regulations.

We previously surveyed primary care physicians about their clinical approach to patients with a possible uncomplicated urinary tract infection (UTI).¹⁵ That 1994-1995 survey included primary care physicians in 4 states and contained detailed information on their work settings and clinical behavior concerning cases of UTI. In this paper we use those survey results to examine the potential effects of office laboratory quality regulations and test availability on testing for UTIs. This is possible because one of the states surveyed, Pennsylvania, has had statewide physician office laboratory quality regulations similar to CLIA-88 in place since the mid-1970s.^12,16 All office laboratories in Pennsylvania were required to register with the state, have a written procedure manual, perform quality and proficiency testing, and comply with state regulations, regardless of the level of testing.¹⁶ Office laboratories in the other 3 states were not regulated until the implementation of CLIA-88 in late 1994. Thus, the survey includes 1 regulated state and 3 states that were just beginning to be regulated. This should provide a conservative estimate of the differences between a regulated and a nonregulated situation, since some of the physicians in the previously unregulated states may already have adjusted to CLIA-88 when our survey was done. The adjustment to CLIA-88 did not occur instantaneously; laboratories have continued to move toward a higher percentage of waived and provider-performed microscopy laboratories (from 57.8% of all laboratories in 1995 to 75% in July 2000) since CLIA-88 was implemented.¹⁷ We sought to determine if office laboratory testing capabilities or availability differs by state and if self-reported use of the tests in diagnosing an uncomplicated UTI is related to the availability of the tests.

Methods

We examined the results of an 88-item survey concerning each physician’s clinical approach to patients with a possible UTI. We mailed the survey in 1994 and 1995 to practicing primary care physicians in 4 specialties that commonly treat adults with uncomplicated UTIs: general internal medicine, family practice, obstetrics and gynecology, and emergency medicine. The emergency medicine physicians were excluded from this analysis, since most of these physicians are hospital based and do not have an office laboratory in the same sense as the other specialties. To obtain a geographically diverse sample, we surveyed physicians from 4 states: Alabama, Nebraska, Minnesota, and Pennsylvania. The names and addresses of potential respondents were obtained from the physician masterfile of the American Medical Association (AMA). Since the AMA masterfile has limited data on physician characteristics (eg, the physician’s specialty), all data in our analysis were based on self-reported survey results. We surveyed the entire population of eligible physicians in each state except Pennsylvania; a random sample was chosen from Pennsylvania because of the large number of practicing physicians in that state.

The survey asked detailed questions about each physician’s clinical approach to a 30-year-old woman with dysuria, a presentation suggestive but not diagnostic of a UTI.^18,19 The main outcomes included whether specific tests (urine dipstick, microscopic urinalysis [UA], wet prep, and urine culture) were available in the office and whether these tests were used when diagnosing a UTI in the given patient. In both parts of the analysis we controlled for possible confounding variables. These included items reflecting the physician’s belief in the usefulness of clinical and laboratory information in diagnosing UTIs and physician and practice characteristics. The variables used in the analyses are shown in Table 1.

In the first part of the analysis we determined whether practicing in a regulated state (Pennsylvania) is associated with changes in the likelihood of having tests available in the office. The 4 outcome variables in this part of the analysis included the presence or absence of the dipstick, microscopic UA, wet prep, and urine culture. Thus, test availability was analyzed using binary dependent variables (yes=the physician reported the test in the office; no=otherwise). We did multivariate logistic regressions because of the binary nature of the dependent variables. Variables explaining the presence or absence of the test in the office included the key variable of interest: the physician’s state of residence (either in Pennsylvania or 1 of the previously unregulated states) and the group of variables reflecting the clinical beliefs concerning history, physical examination, and test usefulness in diagnosing UTI in our hypothetical patient, as well as physician, practice, and community characteristics. The explanatory variables were included in the regression models to control for any factors that might confound the effects of being from Pennsylvania. We hypothesized that tests would be found less frequently in Pennsylvania when controlling for other factors.

In the second part of the analysis, we examined whether the availability of tests in the office is related to their use in diagnosing UTIs in the hypothetical patient. The outcome variables in the second part were the self-reported frequency of ordering microscopic UA, urine culture, and the urine dipstick test (the leukocyte esterase and/or the nitrite test). Test use was analyzed as a binary variable (yes=the physician sometimes or usually performed the test; no=the physician rarely did the test). We did not have information on the physician’s use of the wet prep test, so we could not analyze the relationship between availablity and use for the wet prep. Because we analyzed the variable representing test use as a binary variable, we also used logistic regression for the analysis in the second part. In each regression, the variables explaining the frequency of using the test included 3 outcome measures from the first part of the analysis (whether the dipstick, micro UA, or culture was available in the office), and the same control variables reflecting physician beliefs and the personal and practice characteristics used in the first part of the analysis. We hypothesized that test availability would be related to its use in diagnosing UTIs after controlling for other relevant factors.

We used the Stata statistical software program²⁰ for the analysis.

Results

A total of 8942 surveys were sent out. There were 2172 usable surveys returned. After excluding the responses from the 274 emergency medicine physicians, we analyzed the remaining 1898 responses. We were able to compare respondents and nonrespondents on 3 demographic characteristics: sex, board-certification status, and length of time since graduating from medical school. The survey responders were more likely to be board certified but otherwise were similar to nonrespondents Table 2. Comparison of response patterns in Pennsylvania and the other 3 states using logistic regression indicates that Pennsylvania respondents were less likely to be men than respondents from other states. The magnitude of these differences, however, appears small.

Analysis of Test Availability

The reported prevalence of the various diagnostic tests in office laboratories is shown in Table 3. The dipstick is the most common test; more than 90% of respondents reported having it in their offices. The culture was the least available test, and the prevalence of both the microscopic UA and wet prep were intermediate between the culture and the dipstick. Simple unadjusted comparisons indicate that for each test physicians’ offices in Pennsylvania were less likely to report that they have the test available than were physicians in the unregulated states (P <.0001).

Table 4 shows the determinants of test availability from the multivariate logistic regressions. After controlling for possibly confounding factors, physicians from Pennsylvania were much less likely to have each of the tests in the office. The odds ratios (ORs) range from 0.20 (for the microscopic UA) to 0.35 (for the dipstick). Each of these results is statistically significant.

A number of control variables were significantly related to test availability.* Important findings include the effects of the physician’s specialty, clinical beliefs, and the estimated number of patients with dysuria seen per week. General internal medicine physicians were less likely than family practitioners to have each of the tests in the office (OR = 0.14, 0.60, 0.35, 0.57 for the dipstick, micro UA, wet prep, and culture; all P values <.05), while obstetrician-gynecologists were significantly less likely than family practitioners to have a dipstick, micro UA, and culture (OR = 0.14, 0.24, 0.44; all P values <.05) but more likely to have a wet prep available (OR=5.46; P=.0001). Increased belief in the importance of the leukocyte eserase test for diagnosing UTIs was associated with an increase in dipstick availability (OR=3.40; P=.0001), and increased belief in the importance of the microscopic white cell and red cell readings for diagnosing UTIs was associated with an increase in micro UA availability (OR = 1.95, 1.83; P = .01, .001, respectively). Seeing an additional patient per week with dysuria was associated with an increased availability for the dipstick (OR=1.12; P=.04), micro UA (OR=1.06; P=.01), and wet prep (OR=1.06; P=.02).

The overall explanatory power of these models was fairly good. The C-statistic, representing the area under the receiver operating characteristic curve, ranged from 0.75 for the urine culture to 0.93 for the dipstick (a C-statistic of 0.5 indicates that the model is no better than random chance at predicting the outcome, while a C-statistic of 1 indicates perfect discriminating ability).

Test Availability and the Diagnostic Approach

The relationship between test availability and use demonstrates that the availability of some tests is associated with an increase in their use, while sometimes the availability of tests is associated with an increase or decrease in the use of other tests Table 5. That is, tests will sometimes substitute for or complement other tests when diagnosing UTIs. Physicians with a microscopic UA in the office were more likely to report ordering a microscopic UA. The presence of the urine culture increased microscopic UA use, but the dipstick, when available, appeared to substitute for the microscopic UA.

The reported frequency of ordering urine cultures was increased by the availability of a culture but was decreased when respondents reported having a microscopic UA or dipstick available.

The dipstick was the only test for which use was not significantly related to the presence of any of the diagnostic tests; the availability of the dipstick appears to increase its use, but this relationship was not significant.

Even though test availability is often related to use, the physician may not be the person actually doing the tests. Of the 1492 physicians who report sometimes or usually ordering microscopic UAs, 363 do the procedure themselves, and 803 use ancillary personnel in their office. The remainder (326) have the procedure done outside the office. In contrast, only 25 of the physicians who report sometimes or usually ordering a urine culture do the test themselves. Three hundred forty-two use ancillary help in the office, and 243 have the culture done at an outside laboratory.

Several physician and practice characteristics were significantly related to the frequency of ordering each of the diagnostic tests. Being an obstetrician-gynecologist relative to family practitioner decreased microscopic UA use (OR=0.41; P=.0001), while a greater belief in the value of the leukocyte esterase test decreased microscopic UA use (OR=0.70; P=.01), and greater belief in the microscopic white cell reading increased use of the microscopic UA (OR=3.12; P=.0001). Increased use of urine cultures was associated with being in a city with a population greater than 100,000 relative to a town of less than 10,000 (OR=1.54; P=.02) and being a government employee relative to being in private practice (OR=9.31; P=.0001), while culture use was decreased in practice sizes of more than 25 physicians relative to being in solo practice (OR=0.39; P=.004). The dipstick was used more commonly when physicians had a greater belief in the value of the leukocyte esterase (OR=5.15; P=.0001) and nitrite (OR=2.68; P=.0001) results. Interestingly, the physician’s specialty was generally not significantly related to test use (apart from the difference between obstetrician-gynecologists and family practitioners in micro UA use), in contrast to the consistent specialty differences found for test availability.

The overall power of each of these models to explain the frequency of test use was moderately good. The C-statistics ranged from 0.69 for ordering a urine culture to 0.82 for ordering the dipstick.

Discussion

This study has 2 main findings. First, for each of the office laboratory tests examined, we found significant differences in test availability between unregulated states and Pennsylvania, a state with a long-standing physician office laboratory regulatory system. The overall pattern is one of diminished test availability in Pennsylvania, after controlling for other explanatory influences. This finding was present for every test we examined, regardless of the cost of performing the test or the level at which it was regulated. This supports the hypothesis that tests are available less often in a regulated state.

The second main finding is that the physician’s diagnostic approach to UTIs was related to the in-office availability of tests. This finding supports our second hypothesis. However, the relationship between test and availability and use was more complex than suspected. Not only was the self-reported use of some tests (microscopic UA and culture) related to their availability, sometimes the availability of a test was associated with a change in the self-reported frequency of using other tests. Thus, some tests appear to substitute for others (availability of the dipstick reduces use of microscopic UAs and cultures, and availability of the microscopic UA decreases culture ordering), while the availability of the culture was associated with increased use of the microscopic UA. We also found significant relationships between the physician’s belief in the usefulness of test results in diagnosing UTI with both the availability and subsequent use of the diagnostic tests.

The multivariate analyses indicated that physician and practice characteristics were related to the availability and use of diagnostic tests. This finding is consistent with previous studies reporting numerous factors influencing the use of laboratory tests.^21-23 To our knowledge, the association between physician belief in the value of the test in diagnosing a condition and test availability and use has not been documented previously. Thus, the influence of quality regulations on test prevalence and use will be modulated by the configuration of these other factors.

This study differs from previous studies on laboratory use in that we have been able to control for many relevant explanatory variables, including the physician’s belief in the diagnostic importance of elements from the patient’s history, physical examination, and test results. Many of these items add significant predictive power in a “logical” direction. For example, the availability of many tests was related to the physician’s belief in the value of the results provided by that test, and these beliefs were related to the physician’s diagnostic approach. Although it is not possible to tell from the survey if the physician’s belief in the usefulness of the test is causally related to the availability of a test, the results indicate the need to control for physician-specific opinions and beliefs when examining laboratory test use.

It is not possible to determine the net effect of quality regulations on patient welfare from this study. In theory, a reduction in test availability and subsequent use of a less accurate diagnostic approach is offset by higher quality and accuracy in those offices still offering tests.²⁴ We did not measure the quality or accuracy of the physician’s laboratory testing. To the extent that poor-quality laboratories are closed, regulations should improve overall welfare. Alternatively, if the reduction in test availability occurs in both good- and bad-quality laboratories, then the net effects on patient outcomes may be negative. A recent cost-effectiveness study of tests used to diagnose UTIs²⁵ indicated that the microscopic UA was reasonably cost-effective ($2964 per quality-adjusted life month), but the low sensitivity of the dipstick resulted in a poor cost-effectiveness ratio ($48,460 per quality-adjusted life month). Thus, to the extent that the highly prevalent dipstick is substituting for more cost-effective tests, as our results suggest, regulations might lead to less cost-effective care.

Our analysis suggests that changing the availability of tests may change physicians’ diagnostic approaches in ways that are difficult to predict a priori. For example, reducing the availability of the microscopic UA might lead to greater use of the more expensive urine culture. When many tests are reduced in availability, it is difficult to anticipate how the care of patients will change, since tests will substitute for or complement the use of other tests in ways that are difficult to predict.

Limitations

This study has several limitations. First, it is based on survey data that may not reflect actual behavior. However, the results are plausible, internally consistent, and consistent with previous research findings documenting substantial variation in practice patterns in managing UTIs.²⁶ Our focus on how physicians treat an uncomplicated basic UTI limits the possibility of eliciting responses based on unmeasured patient characteristics. Second, the overall response rate was low, which might raise concerns that the respondents represent an atypical or biased group of physicians. A comparison of respondents and nonrespondents along the 3 characteristics we had access to did not reveal any striking differences. And, when compared with nationwide physician work force statistics collected by the AMA in 1994, the physicians in this study had similar hours per week spent in patient care (46.2 in our survey vs 48.2 in the AMA survey) and estimated level of capitation (19.3% in our survey vs 16.7% in the AMA survey).²⁷ The estimated volume of patients with UTI was also similar to previous surveys of primary care physicians (approximately 6 per week).²⁶ Thus, the respondents in this survey do not seem to represent an atypical or unusual group of physicians. Other limitations include the fact that physicians responding to the survey might not be responsible for deciding whether to acquire or maintain tests in the office. Likewise, the survey does not contain information on other factors in the physician’s local market that might influence the supply or demand for office tests. These effects, however, would have to be quite large and disproportionately affect Pennsylvania respondents to confound our findings.

Finally, since the analysis is based on cross-sectional data, the results do not give any indication of how rapidly in-office testing declines after the implementation of quality regulations. However, as mentioned before, statistics indicate that the mix of tests in physicians’ offices has continued to move towards simpler tests since CLIA-88 was implemented. Our study provides a snapshot of the circumstances as CLIA-88 was first implemented; it would be worthwhile to assess changes in test prevalence and use in our survey population over the last several years. It may be the case that other changes underway in health care delivery (eg, improving technology, vertical and horizontal integration of physician groups, and growth of managed care) may hasten or reverse the changes in office laboratory availability implied in this study.

Conclusions

The availability of common office tests used in the diagnosis of UTI appears to be related to many factors, including the presence of office laboratory regulations. Lower availability of tests, in turn, was associated with the diagnostic approach when treating patients with a possible UTI. The net effects on patient welfare are difficult to predict, but our findings raise a serious concern about the possibility of a detrimental overall effect on patient welfare. Determining the full effect of quality regulations will require careful study to measure the changes in treatment patterns, treatment quality, and patient outcomes subsequent to the implementation of increased levels of regulation.

Acknowledgments

This research was funded by the Department of Internal Medicine at the University of Nebraska Medical Center, Omaha, and by the Department of Internal Medicine at Abington Memorial Hospital, Philadelphia, Pennsylvania.

Busy clinicians have come to rely on the office laboratory as an important tool in providing timely, efficient, and high-quality patient care. However, concerns have been raised about poor accuracy in physician office laboratories and financial incentives that encourage overuse of tests.^1-7 Recent studies have documented inaccurate office laboratory testing.^8,9 Previously, the Commission on Office Laboratory Accreditation concluded that approximately 5% of office laboratories have serious deficiencies involving quality control, instrument maintenance, specimen management, and proficiency testing.¹⁰ Similar concerns led to the passage of the Clinical Laboratory Improvement Act of 1988 (CLIA-88), which was implemented in the fall of 1994.¹¹ Before CLIA-88, physician office laboratory quality regulations were decided at the state level.¹² Both state and federal regulations require adherence to specific measures (eg, proficiency testing, quality control, and quality assurance) to operate an office laboratory.

The critics of quality regulations assert that the measures might actually decrease the quality of care. They note that the regulations impose a fixed cost on the physician’s practice that must be spent regardless of how busy the laboratory is. The additional costs of compliance could discourage physicians from offering tests that would otherwise be done.^13,14 This reduction in test availability might occur in both high- and low-quality laboratories. Therefore, regulations aimed at improving quality might have the unintended effect of reducing the availability and use of clinically beneficial laboratory tests, thereby decreasing the overall potential benefit of the regulations.

We previously surveyed primary care physicians about their clinical approach to patients with a possible uncomplicated urinary tract infection (UTI).¹⁵ That 1994-1995 survey included primary care physicians in 4 states and contained detailed information on their work settings and clinical behavior concerning cases of UTI. In this paper we use those survey results to examine the potential effects of office laboratory quality regulations and test availability on testing for UTIs. This is possible because one of the states surveyed, Pennsylvania, has had statewide physician office laboratory quality regulations similar to CLIA-88 in place since the mid-1970s.^12,16 All office laboratories in Pennsylvania were required to register with the state, have a written procedure manual, perform quality and proficiency testing, and comply with state regulations, regardless of the level of testing.¹⁶ Office laboratories in the other 3 states were not regulated until the implementation of CLIA-88 in late 1994. Thus, the survey includes 1 regulated state and 3 states that were just beginning to be regulated. This should provide a conservative estimate of the differences between a regulated and a nonregulated situation, since some of the physicians in the previously unregulated states may already have adjusted to CLIA-88 when our survey was done. The adjustment to CLIA-88 did not occur instantaneously; laboratories have continued to move toward a higher percentage of waived and provider-performed microscopy laboratories (from 57.8% of all laboratories in 1995 to 75% in July 2000) since CLIA-88 was implemented.¹⁷ We sought to determine if office laboratory testing capabilities or availability differs by state and if self-reported use of the tests in diagnosing an uncomplicated UTI is related to the availability of the tests.

Methods

We examined the results of an 88-item survey concerning each physician’s clinical approach to patients with a possible UTI. We mailed the survey in 1994 and 1995 to practicing primary care physicians in 4 specialties that commonly treat adults with uncomplicated UTIs: general internal medicine, family practice, obstetrics and gynecology, and emergency medicine. The emergency medicine physicians were excluded from this analysis, since most of these physicians are hospital based and do not have an office laboratory in the same sense as the other specialties. To obtain a geographically diverse sample, we surveyed physicians from 4 states: Alabama, Nebraska, Minnesota, and Pennsylvania. The names and addresses of potential respondents were obtained from the physician masterfile of the American Medical Association (AMA). Since the AMA masterfile has limited data on physician characteristics (eg, the physician’s specialty), all data in our analysis were based on self-reported survey results. We surveyed the entire population of eligible physicians in each state except Pennsylvania; a random sample was chosen from Pennsylvania because of the large number of practicing physicians in that state.

The survey asked detailed questions about each physician’s clinical approach to a 30-year-old woman with dysuria, a presentation suggestive but not diagnostic of a UTI.^18,19 The main outcomes included whether specific tests (urine dipstick, microscopic urinalysis [UA], wet prep, and urine culture) were available in the office and whether these tests were used when diagnosing a UTI in the given patient. In both parts of the analysis we controlled for possible confounding variables. These included items reflecting the physician’s belief in the usefulness of clinical and laboratory information in diagnosing UTIs and physician and practice characteristics. The variables used in the analyses are shown in Table 1.

In the first part of the analysis we determined whether practicing in a regulated state (Pennsylvania) is associated with changes in the likelihood of having tests available in the office. The 4 outcome variables in this part of the analysis included the presence or absence of the dipstick, microscopic UA, wet prep, and urine culture. Thus, test availability was analyzed using binary dependent variables (yes=the physician reported the test in the office; no=otherwise). We did multivariate logistic regressions because of the binary nature of the dependent variables. Variables explaining the presence or absence of the test in the office included the key variable of interest: the physician’s state of residence (either in Pennsylvania or 1 of the previously unregulated states) and the group of variables reflecting the clinical beliefs concerning history, physical examination, and test usefulness in diagnosing UTI in our hypothetical patient, as well as physician, practice, and community characteristics. The explanatory variables were included in the regression models to control for any factors that might confound the effects of being from Pennsylvania. We hypothesized that tests would be found less frequently in Pennsylvania when controlling for other factors.

In the second part of the analysis, we examined whether the availability of tests in the office is related to their use in diagnosing UTIs in the hypothetical patient. The outcome variables in the second part were the self-reported frequency of ordering microscopic UA, urine culture, and the urine dipstick test (the leukocyte esterase and/or the nitrite test). Test use was analyzed as a binary variable (yes=the physician sometimes or usually performed the test; no=the physician rarely did the test). We did not have information on the physician’s use of the wet prep test, so we could not analyze the relationship between availablity and use for the wet prep. Because we analyzed the variable representing test use as a binary variable, we also used logistic regression for the analysis in the second part. In each regression, the variables explaining the frequency of using the test included 3 outcome measures from the first part of the analysis (whether the dipstick, micro UA, or culture was available in the office), and the same control variables reflecting physician beliefs and the personal and practice characteristics used in the first part of the analysis. We hypothesized that test availability would be related to its use in diagnosing UTIs after controlling for other relevant factors.

We used the Stata statistical software program²⁰ for the analysis.

Results

A total of 8942 surveys were sent out. There were 2172 usable surveys returned. After excluding the responses from the 274 emergency medicine physicians, we analyzed the remaining 1898 responses. We were able to compare respondents and nonrespondents on 3 demographic characteristics: sex, board-certification status, and length of time since graduating from medical school. The survey responders were more likely to be board certified but otherwise were similar to nonrespondents Table 2. Comparison of response patterns in Pennsylvania and the other 3 states using logistic regression indicates that Pennsylvania respondents were less likely to be men than respondents from other states. The magnitude of these differences, however, appears small.

Analysis of Test Availability

The reported prevalence of the various diagnostic tests in office laboratories is shown in Table 3. The dipstick is the most common test; more than 90% of respondents reported having it in their offices. The culture was the least available test, and the prevalence of both the microscopic UA and wet prep were intermediate between the culture and the dipstick. Simple unadjusted comparisons indicate that for each test physicians’ offices in Pennsylvania were less likely to report that they have the test available than were physicians in the unregulated states (P <.0001).

Table 4 shows the determinants of test availability from the multivariate logistic regressions. After controlling for possibly confounding factors, physicians from Pennsylvania were much less likely to have each of the tests in the office. The odds ratios (ORs) range from 0.20 (for the microscopic UA) to 0.35 (for the dipstick). Each of these results is statistically significant.

A number of control variables were significantly related to test availability.* Important findings include the effects of the physician’s specialty, clinical beliefs, and the estimated number of patients with dysuria seen per week. General internal medicine physicians were less likely than family practitioners to have each of the tests in the office (OR = 0.14, 0.60, 0.35, 0.57 for the dipstick, micro UA, wet prep, and culture; all P values <.05), while obstetrician-gynecologists were significantly less likely than family practitioners to have a dipstick, micro UA, and culture (OR = 0.14, 0.24, 0.44; all P values <.05) but more likely to have a wet prep available (OR=5.46; P=.0001). Increased belief in the importance of the leukocyte eserase test for diagnosing UTIs was associated with an increase in dipstick availability (OR=3.40; P=.0001), and increased belief in the importance of the microscopic white cell and red cell readings for diagnosing UTIs was associated with an increase in micro UA availability (OR = 1.95, 1.83; P = .01, .001, respectively). Seeing an additional patient per week with dysuria was associated with an increased availability for the dipstick (OR=1.12; P=.04), micro UA (OR=1.06; P=.01), and wet prep (OR=1.06; P=.02).

The overall explanatory power of these models was fairly good. The C-statistic, representing the area under the receiver operating characteristic curve, ranged from 0.75 for the urine culture to 0.93 for the dipstick (a C-statistic of 0.5 indicates that the model is no better than random chance at predicting the outcome, while a C-statistic of 1 indicates perfect discriminating ability).

Test Availability and the Diagnostic Approach

The relationship between test availability and use demonstrates that the availability of some tests is associated with an increase in their use, while sometimes the availability of tests is associated with an increase or decrease in the use of other tests Table 5. That is, tests will sometimes substitute for or complement other tests when diagnosing UTIs. Physicians with a microscopic UA in the office were more likely to report ordering a microscopic UA. The presence of the urine culture increased microscopic UA use, but the dipstick, when available, appeared to substitute for the microscopic UA.

The reported frequency of ordering urine cultures was increased by the availability of a culture but was decreased when respondents reported having a microscopic UA or dipstick available.

The dipstick was the only test for which use was not significantly related to the presence of any of the diagnostic tests; the availability of the dipstick appears to increase its use, but this relationship was not significant.

Even though test availability is often related to use, the physician may not be the person actually doing the tests. Of the 1492 physicians who report sometimes or usually ordering microscopic UAs, 363 do the procedure themselves, and 803 use ancillary personnel in their office. The remainder (326) have the procedure done outside the office. In contrast, only 25 of the physicians who report sometimes or usually ordering a urine culture do the test themselves. Three hundred forty-two use ancillary help in the office, and 243 have the culture done at an outside laboratory.

Several physician and practice characteristics were significantly related to the frequency of ordering each of the diagnostic tests. Being an obstetrician-gynecologist relative to family practitioner decreased microscopic UA use (OR=0.41; P=.0001), while a greater belief in the value of the leukocyte esterase test decreased microscopic UA use (OR=0.70; P=.01), and greater belief in the microscopic white cell reading increased use of the microscopic UA (OR=3.12; P=.0001). Increased use of urine cultures was associated with being in a city with a population greater than 100,000 relative to a town of less than 10,000 (OR=1.54; P=.02) and being a government employee relative to being in private practice (OR=9.31; P=.0001), while culture use was decreased in practice sizes of more than 25 physicians relative to being in solo practice (OR=0.39; P=.004). The dipstick was used more commonly when physicians had a greater belief in the value of the leukocyte esterase (OR=5.15; P=.0001) and nitrite (OR=2.68; P=.0001) results. Interestingly, the physician’s specialty was generally not significantly related to test use (apart from the difference between obstetrician-gynecologists and family practitioners in micro UA use), in contrast to the consistent specialty differences found for test availability.

The overall power of each of these models to explain the frequency of test use was moderately good. The C-statistics ranged from 0.69 for ordering a urine culture to 0.82 for ordering the dipstick.

Discussion

This study has 2 main findings. First, for each of the office laboratory tests examined, we found significant differences in test availability between unregulated states and Pennsylvania, a state with a long-standing physician office laboratory regulatory system. The overall pattern is one of diminished test availability in Pennsylvania, after controlling for other explanatory influences. This finding was present for every test we examined, regardless of the cost of performing the test or the level at which it was regulated. This supports the hypothesis that tests are available less often in a regulated state.

The second main finding is that the physician’s diagnostic approach to UTIs was related to the in-office availability of tests. This finding supports our second hypothesis. However, the relationship between test and availability and use was more complex than suspected. Not only was the self-reported use of some tests (microscopic UA and culture) related to their availability, sometimes the availability of a test was associated with a change in the self-reported frequency of using other tests. Thus, some tests appear to substitute for others (availability of the dipstick reduces use of microscopic UAs and cultures, and availability of the microscopic UA decreases culture ordering), while the availability of the culture was associated with increased use of the microscopic UA. We also found significant relationships between the physician’s belief in the usefulness of test results in diagnosing UTI with both the availability and subsequent use of the diagnostic tests.

The multivariate analyses indicated that physician and practice characteristics were related to the availability and use of diagnostic tests. This finding is consistent with previous studies reporting numerous factors influencing the use of laboratory tests.^21-23 To our knowledge, the association between physician belief in the value of the test in diagnosing a condition and test availability and use has not been documented previously. Thus, the influence of quality regulations on test prevalence and use will be modulated by the configuration of these other factors.

This study differs from previous studies on laboratory use in that we have been able to control for many relevant explanatory variables, including the physician’s belief in the diagnostic importance of elements from the patient’s history, physical examination, and test results. Many of these items add significant predictive power in a “logical” direction. For example, the availability of many tests was related to the physician’s belief in the value of the results provided by that test, and these beliefs were related to the physician’s diagnostic approach. Although it is not possible to tell from the survey if the physician’s belief in the usefulness of the test is causally related to the availability of a test, the results indicate the need to control for physician-specific opinions and beliefs when examining laboratory test use.

It is not possible to determine the net effect of quality regulations on patient welfare from this study. In theory, a reduction in test availability and subsequent use of a less accurate diagnostic approach is offset by higher quality and accuracy in those offices still offering tests.²⁴ We did not measure the quality or accuracy of the physician’s laboratory testing. To the extent that poor-quality laboratories are closed, regulations should improve overall welfare. Alternatively, if the reduction in test availability occurs in both good- and bad-quality laboratories, then the net effects on patient outcomes may be negative. A recent cost-effectiveness study of tests used to diagnose UTIs²⁵ indicated that the microscopic UA was reasonably cost-effective ($2964 per quality-adjusted life month), but the low sensitivity of the dipstick resulted in a poor cost-effectiveness ratio ($48,460 per quality-adjusted life month). Thus, to the extent that the highly prevalent dipstick is substituting for more cost-effective tests, as our results suggest, regulations might lead to less cost-effective care.

Our analysis suggests that changing the availability of tests may change physicians’ diagnostic approaches in ways that are difficult to predict a priori. For example, reducing the availability of the microscopic UA might lead to greater use of the more expensive urine culture. When many tests are reduced in availability, it is difficult to anticipate how the care of patients will change, since tests will substitute for or complement the use of other tests in ways that are difficult to predict.

Limitations

This study has several limitations. First, it is based on survey data that may not reflect actual behavior. However, the results are plausible, internally consistent, and consistent with previous research findings documenting substantial variation in practice patterns in managing UTIs.²⁶ Our focus on how physicians treat an uncomplicated basic UTI limits the possibility of eliciting responses based on unmeasured patient characteristics. Second, the overall response rate was low, which might raise concerns that the respondents represent an atypical or biased group of physicians. A comparison of respondents and nonrespondents along the 3 characteristics we had access to did not reveal any striking differences. And, when compared with nationwide physician work force statistics collected by the AMA in 1994, the physicians in this study had similar hours per week spent in patient care (46.2 in our survey vs 48.2 in the AMA survey) and estimated level of capitation (19.3% in our survey vs 16.7% in the AMA survey).²⁷ The estimated volume of patients with UTI was also similar to previous surveys of primary care physicians (approximately 6 per week).²⁶ Thus, the respondents in this survey do not seem to represent an atypical or unusual group of physicians. Other limitations include the fact that physicians responding to the survey might not be responsible for deciding whether to acquire or maintain tests in the office. Likewise, the survey does not contain information on other factors in the physician’s local market that might influence the supply or demand for office tests. These effects, however, would have to be quite large and disproportionately affect Pennsylvania respondents to confound our findings.

Finally, since the analysis is based on cross-sectional data, the results do not give any indication of how rapidly in-office testing declines after the implementation of quality regulations. However, as mentioned before, statistics indicate that the mix of tests in physicians’ offices has continued to move towards simpler tests since CLIA-88 was implemented. Our study provides a snapshot of the circumstances as CLIA-88 was first implemented; it would be worthwhile to assess changes in test prevalence and use in our survey population over the last several years. It may be the case that other changes underway in health care delivery (eg, improving technology, vertical and horizontal integration of physician groups, and growth of managed care) may hasten or reverse the changes in office laboratory availability implied in this study.

Conclusions

The availability of common office tests used in the diagnosis of UTI appears to be related to many factors, including the presence of office laboratory regulations. Lower availability of tests, in turn, was associated with the diagnostic approach when treating patients with a possible UTI. The net effects on patient welfare are difficult to predict, but our findings raise a serious concern about the possibility of a detrimental overall effect on patient welfare. Determining the full effect of quality regulations will require careful study to measure the changes in treatment patterns, treatment quality, and patient outcomes subsequent to the implementation of increased levels of regulation.

Acknowledgments

This research was funded by the Department of Internal Medicine at the University of Nebraska Medical Center, Omaha, and by the Department of Internal Medicine at Abington Memorial Hospital, Philadelphia, Pennsylvania.

Approximately 50 million Papanicolaou (Pap) tests are performed annually in the United States, of which approximately 5% will require further evaluation because of an abnormality.¹ The Bethesda System for reporting cervical/vaginal cytologic diagnoses,² developed in 1988 and revised in 1991, provides uniform diagnostic terminology to improve consistency of reporting. The Bethesda System introduced new terminology for atypical cytology, creating the label “atypical squamous cells of uncertain significance” (ASCUS). Pap test results labelled ASCUS encompass a spectrum of cellular change reflecting a variety of pathologic processes that cannot be more specifically categorized, including reactive changes, inflammation, human papillomavirus (HPV) related changes, suggested dysplasia, and less than optimal slide preparation.² Because studies have consistently shown a small but real risk that an ASCUS Pap result represents an underlying high-grade lesion,^3-6 there has been much debate about their management. Uncertainty about optimal management led to the development of clinical practice guidelines (CPG).^7-9

The National Cancer Institute⁷ (NCI) and a Canadian expert pane¹⁸ suggest that ASCUS and low-grade squamous intraepithelial lesions can be managed either by immediate colposcopy or by repeating cervical cytology. Clinicians may, if they choose, reserve referral to colposcopy for those patients with persistent abnormality on subsequent cytology. This conservative strategy is based on several assumptions:¹⁰ (1) conditions that mimic dysplasia, including reactive changes, are common; (2) spontaneous regression occurs in at least 50% of low-grade lesions; (3) the rate of progression is relatively slow; (4) the chance of carcinoma in situ with an ASCUS Pap test result is low; and (5) persistently negative follow-up Pap results are unlikely to be false negatives. The current practice guideline suggests that if clinicians opt for cytologic follow-up of ASCUS, repeat Pap tests should be done every 4 to 6 months until 3 consecutive normal test results have been obtained. Colposcopy should be performed if there is a subsequent abnormal Pap test result.⁷ The 1996 statement of the American College of Obstetricians and Gynecologists is similar, with the addition that colposcopy be performed for women with a single ASCUS Pap result if they also have any of the following high-risk factors: HPV infection, human immunodeficiency virus (HIV) infection, smoking, or multiple sexual partners.⁹

A disproportionate share of the cervical cancer burden in this country, in terms of both incidence and mortality, is shouldered by low-income minority women.^11-13 Minority and low-income populations have the highest rates of nonadherence with repeat Pap test or colposcopy.^14-18 Thus, the women who are at the highest risk of cervical intraepithelial neoplasia (CIN) may also be the most at risk for suboptimal management after an abnormal Pap test result is obtained. This project assessed current management of ASCUS Pap results in a diverse community at high risk of cervical cancer by comparing actual practice to the NCI guideline. In addition to assessing levels of follow-up in this community, we sought to identify predictors of adherence to that guideline.

Methods

Setting and Population

Data were collected in 7 community health centers serving a low-income, predominantly minority population. The centers represent diversity with regard to practice type (medicine/pediatrics/obstetrics model or family medicine model) and size (20,000-80,000 visits/year). The subjects were women with newly identified atypia or ASCUS Pap test results in 1996; women for whom the index Pap test represented persistent atypia or ASCUS were excluded, as were women with a history of CIN or squamous intraepithelial lesion (SIL). Those who were known to be infected with HIV represent a special category^19,20 that may have been managed differently at the clinician’s discretion; therefore, they were also excluded from the analysis.

Data Collection

After approval by the appropriate institutional review boards, women with atypical results in 1996 were identified by 2 mechanisms: site logs maintained for clinical tracking, and because all sites exclusively use the same laboratory, the central cytology database. Pap test results described as atypia or ASCUS were included, as these represent different terminology for the same spectrum of cytologic abnormalities with continued use of the older term by some cytopathologists. Chart abstraction by a trained research assistant collected information regarding: (1) complicating medical conditions (immunosuppression, history of cancer, substance abuse, pregnancy); (2) history of cervical dysplasia; (3) primary care provider and number of visits to site; (4) the provider’s management plan; (5) all subsequent Pap test and colposcopy results; (6) evidence of communication of the management plan to the subject; and (7) demographics.

Accuracy of chart abstraction was confirmed by the lead author rescreening 20% of the records. The data were entered into a Paradox database (Borland Software Corporation, Scotts Valley, Calif), and statistical analysis was performed using SAS software (SAS Institute, Inc, Cary, NC) and Stata software (Stata Corporation, College Station, Tex).

Analysis

The goal of our analysis was to determine the factors associated with follow-up. We defined 3 mutually exclusive levels of adherence to the CPG: (1) complete adherence (colposcopy done if recommended or 3 repeat tests approximately every 6 months until 3 are normal or colposcopy if there is persistent abnormality); (2) moderate adherence (1 or 2 follow-up tests after the index Pap test); and (3) low adherence (no follow-up cytology or histology, or abnormality on repeat Pap test without colposcopy.)

Univariate analyses were performed relating the level of follow-up to the description of the abnormality (atypia vs ASCUS), patient demographics, and practice and provider characteristics (eg, training, availability of colposcopy on site, and plan of care). In analyses where cell size fell below 5, Fisher exact test or Yates correction (for analyses with row or column size greater than 2) replaced the Pearson chi-square. We performed multivariate analysis using a hierarchical model with patients nested within sites, and ordinal logistic regression to account for the 3 ordered levels of adherence to the CPG. Variables significant at P less than .2 in the univariate analyses were included in the multivariate model to ensure inclusion of potentially important variables. Also, insurance status was included, because others have found significant associations for this variable.^21-23 The provider’s management plan was not included as a variable in the regression model, because it is not sufficiently independent of the adherence levels.

Because subjects were clustered within 7 sites (not a simple random sample), we treated site as a random effect with patients nested within the site. An ordered logistic regression model was fit by maximum likelihood estimation, using the program GLLAMM⁶ in Stata version 6.0 The fit of the ordered logistic regression model to the data was tested using the method of Harrell and colleagues.²⁴ Briefly, this method compares observed mean values of the model’s linear combination of the predictor variables to expected values under the assumption that the model (and in particular the proportional odds ratio assumption) is correctly specified.

Results

of the 570 women identified with an atypical or ASCUS Pap test result in 1996, 554 charts were located and reviewed. The charts of 16 individuals could not be located for review. Some women (n=69) identified had a previous atypical or ASCUS Pap test result. The sample of women with no previous ASCUS or atypical result, excluding those with documented HIV infection (n=19) and those with a history of CIN or SIL (n=79), consisted of 387 subjects. The mean age of the subjects at the time of the index Pap test was 32 years (standard deviation=11.7). More than half of the subjects (58.4%) were Hispanic. The majority of subjects had Medicaid (73.1%). The majority of index Pap result abnormalities were described as atypia (61.5%), with the remainder ASCUS (38.5%). A small percentage of women (8.8%) were pregnant at the time of the index Pap test. The subjects were approximately evenly split between receiving care in medicine/pediatrics/obstetrics model sites and family medicine model sites. Colposcopy was provided on site (at least 2 times per week) in 4 of the centers that served 68.7% of the subjects.

Overall, many subjects remained in care in the health centers for a relatively short time. Almost one fourth of the subjects (23.7%) had a final visit to the health center less than 6 months from the index Pap test. Complete data including 2 full years of follow-up from the date of the index Pap test were available for 41% of subjects. Subjects with new atypia had a small chance of SIL on subsequent tests (6.5%, increasing to 10.9% for those subjects with persistent atypia). Results of biopsies for women with new and persistent atypia are presented in Table 1. The majority of subjects had normal or low-grade findings on histology, and no cancer was diagnosed.

Management of atypical results is depicted in the Figure 1. Providers were credited with a “correct” plan if they recommended either a repeat test in 6 months or less, or colposcopy. “Incorrect” management included any other recommendation or failure to document any plan (18.9%). Providers were most likely to recommend a repeat test in 3 months (31.8%) or 6 months (46.1%). We considered communication to subjects “complete” if the notes reflected discussion of the result at a visit (58.3%) or if a completed telephone call was documented (3.9%). Communication was considered attempted if a letter was sent (47.2%) or a call attempted (8.8%) and documented. In 8.3% of charts no documentation of attempted or completed communication was present. There were no significant associations of provider plan or communication of results with demographic variables, pregnancy, or abnormality type. Family physicians were more likely than midlevel providers or other physicians to generate a correct plan (87.3%, 80.8%, 69.8%, respectively; P=.002). Family physicians documented completed communication of results in 77.8% of cases, compared with 70.1% for other physicians and 57.4% for midlevel providers.

Providers were more likely to recommend colposcopy when the index Pap test result was reported as ASCUS rather than atypia (22.1% vs 11.4%, P=.016). Providers were much more likely to choose colposcopy if the pathologist made a recommendation for colposcopy rather than a repeat test (85.2% vs 8.9%, P=.001) and if the report included “favored SIL” in the result (75.0% vs 11.5%, P=.001.) When colposcopy was recommended, 70.8% completed the procedure. After a first atypical Pap test, 36% had no follow-up test; 27% had 1 follow-up test; 20% had 2 tests; and 17% had 3 subsequent tests. Subjects were found to have complete, moderate, and low adherence levels as follows: 26.9%, 27.6%, and 45.5%.

Univariate associations with levels of adherence are shown in Table 2. Complete adherence was more likely when the result was discussed at a visit, when the provider’s plan was documented, when the provider was other than a family physician, and if the plan was colposcopy as opposed to repeat cytology. Complete adherence was associated with remaining in care at the centers after 1997. It was more likely for results reported as ASCUS as opposed to atypia. Large differences are seen in the number of women with complete adherence by site, with the percentage ranging from 7% to 57%.

In the ordered logistic regression analysis, 4 factors were found to be predictive of the quality of adherence: the description of the abnormality (atypia vs ASCUS), availability of colposcopy on-site, discussion of the plan at a visit, and site. The detailed results are shown in Table 3. The adjusted odds ratio for complete versus moderate and low adherence (or equivalently moderate and complete vs low) associated with having an ASCUS lesion versus atypia is 1.56; for attending a clinic where colposcopy is referred off-site, 0.39; and for having discussed the follow-up plan, 3.44. The effects of ethnicity, type of health insurance, provider training, and pregnancy were not statistically significant.

Discussion

The introduction of the ASCUS label in the early 1990s created controversy and led to guidelines from the NCI7 and others^8,9 based primarily on expert opinion. Data from a large randomized clinical trial25 currently underway may resolve some of the questions about the effectiveness of repeat cytology versus immediate colposcopy in the management of low-grade dysplasia. Importantly, lack of evidence for the guideline may adversely affect compliance. Implementation of either strategy is difficult, especially in communities with many risk factors for poor follow-up. Our results indicate low adherence to the NCI interim guideline in an urban primary care setting serving predominately low-income minority women. Of concern is that almost half our sample, compared with 30% of subjects in a Canadian study,²⁶ had no follow-up cytology or histology. Our findings highlight some of the difficulties encountered in managing low-grade abnormalities in a setting challenged by high provider turnover, staffing difficulties, incorrect phone and address information, and financial and language barriers. Long-term repeated cytology is difficult to accomplish where loss to follow-up is common and competing demands of managing multiple chronic illnesses may interfere. Our findings are similar to those of other investigators^27,28 in that colposcopy, when recommended, was not completed by approximately one third of the women.

Our findings highlight the importance of strict use of Bethesda terminology by cytopathologists, since the older terminology may be misunderstood and less aggressively followed by primary care clinicians. We chose to include women with either atypical or ASCUS results, because they are different labels for the same spectrum of abnormalities. We found that providers were slightly more likely to recommend colposcopy for ASCUS and ultimately that women with atypical results were more likely to have low adherence to the CPG. This probably reflects that many primary care providers do not view the guideline as applicable to results reported as atypia or communicate with patients about atypical results differently.

There were substantial differences in rates of follow-up by site that are likely related to several factors. Although tracking of abnormal Pap test results was required at all sites the intensity varied, with the most successful site staffed by registered nurses who worked closely with the colposcopy providers in maintaining a manual card file. Sites where colposcopy was recommended more often had overall higher proportions of women with complete adherence to the CPG than those sites where serial cytology was the follow-up strategy of choice. Family physicians were more likely to generate a correct plan yet had worse overall adherence. This is likely due to a preference to follow up women by serial cytology, which proves difficult in settings where patients often stay in care for less than 2 years. However, the finding persists even after controlling for the provider plan, suggesting that other factors are also contributing to account for the differences between sites. Importantly, providers were more likely to opt for colposcopy, and women were more likely to complete it if the service was available in the subject’s health center.

We are unable to determine all of the causes of inadequate follow-up. Systemic factors, however, are clearly implicated by the present data, as indicated by the substantial variation by site and the number of charts in which a clear provider plan was not indicated or where no indication that the subject had been informed of the result was documented. When documentation was present that results were discussed with patients during visits patients were much more likely to have good follow-up, again suggesting that good communication of results is imperative. Our results also suggest that reducing barriers to colposcopy, by such means as providing the service on-site, may be effective in achieving optimal follow-up when colposcopy is recommended. Recommending colposcopy for all women with an initial ASCUS or atypical Pap test result will result in significant stress on available colposcopic resources and expose large numbers of women to an expensive, uncomfortable, and distressing procedure unnecessarily. However, providers need to carefully consider the risk of loss to follow-up in making a recommendation, take extra effort to insure that women are informed of and understand their Pap test result, and establish clinical tracking efforts to meet the challenge of serial cytology in high-risk settings.

Limitations

Our findings are limited in that the standard of care to which we compared practice is based on a guideline that is not evidence based. Other limitations of the project are largely due to constraints imposed by the abstraction of retrospective data from primary care charts. Information about how women were informed of results is sketchy, and failure to document may result in underestimation of actual patient contact. The extent to which women had follow-up if they no longer received care in the health centers cannot be estimated. Some women may have had follow-up with other providers outside our system. Demographic information was obtained from site registration databases rather than self-report and may include some misclassification. Some important information is not captured consistently in the site’s computers, including changes in insurance and primary care providers. Provider turnover was extremely high in these health centers during the interval studied, though lists of providers are not available to create a precise measure to include in our model.

Acknowledgments

This work was supported by a clinical research training grant from the American Cancer Society.

Approximately 50 million Papanicolaou (Pap) tests are performed annually in the United States, of which approximately 5% will require further evaluation because of an abnormality.¹ The Bethesda System for reporting cervical/vaginal cytologic diagnoses,² developed in 1988 and revised in 1991, provides uniform diagnostic terminology to improve consistency of reporting. The Bethesda System introduced new terminology for atypical cytology, creating the label “atypical squamous cells of uncertain significance” (ASCUS). Pap test results labelled ASCUS encompass a spectrum of cellular change reflecting a variety of pathologic processes that cannot be more specifically categorized, including reactive changes, inflammation, human papillomavirus (HPV) related changes, suggested dysplasia, and less than optimal slide preparation.² Because studies have consistently shown a small but real risk that an ASCUS Pap result represents an underlying high-grade lesion,^3-6 there has been much debate about their management. Uncertainty about optimal management led to the development of clinical practice guidelines (CPG).^7-9

The National Cancer Institute⁷ (NCI) and a Canadian expert pane¹⁸ suggest that ASCUS and low-grade squamous intraepithelial lesions can be managed either by immediate colposcopy or by repeating cervical cytology. Clinicians may, if they choose, reserve referral to colposcopy for those patients with persistent abnormality on subsequent cytology. This conservative strategy is based on several assumptions:¹⁰ (1) conditions that mimic dysplasia, including reactive changes, are common; (2) spontaneous regression occurs in at least 50% of low-grade lesions; (3) the rate of progression is relatively slow; (4) the chance of carcinoma in situ with an ASCUS Pap test result is low; and (5) persistently negative follow-up Pap results are unlikely to be false negatives. The current practice guideline suggests that if clinicians opt for cytologic follow-up of ASCUS, repeat Pap tests should be done every 4 to 6 months until 3 consecutive normal test results have been obtained. Colposcopy should be performed if there is a subsequent abnormal Pap test result.⁷ The 1996 statement of the American College of Obstetricians and Gynecologists is similar, with the addition that colposcopy be performed for women with a single ASCUS Pap result if they also have any of the following high-risk factors: HPV infection, human immunodeficiency virus (HIV) infection, smoking, or multiple sexual partners.⁹

A disproportionate share of the cervical cancer burden in this country, in terms of both incidence and mortality, is shouldered by low-income minority women.^11-13 Minority and low-income populations have the highest rates of nonadherence with repeat Pap test or colposcopy.^14-18 Thus, the women who are at the highest risk of cervical intraepithelial neoplasia (CIN) may also be the most at risk for suboptimal management after an abnormal Pap test result is obtained. This project assessed current management of ASCUS Pap results in a diverse community at high risk of cervical cancer by comparing actual practice to the NCI guideline. In addition to assessing levels of follow-up in this community, we sought to identify predictors of adherence to that guideline.

Methods

Setting and Population

Data were collected in 7 community health centers serving a low-income, predominantly minority population. The centers represent diversity with regard to practice type (medicine/pediatrics/obstetrics model or family medicine model) and size (20,000-80,000 visits/year). The subjects were women with newly identified atypia or ASCUS Pap test results in 1996; women for whom the index Pap test represented persistent atypia or ASCUS were excluded, as were women with a history of CIN or squamous intraepithelial lesion (SIL). Those who were known to be infected with HIV represent a special category^19,20 that may have been managed differently at the clinician’s discretion; therefore, they were also excluded from the analysis.

Data Collection

After approval by the appropriate institutional review boards, women with atypical results in 1996 were identified by 2 mechanisms: site logs maintained for clinical tracking, and because all sites exclusively use the same laboratory, the central cytology database. Pap test results described as atypia or ASCUS were included, as these represent different terminology for the same spectrum of cytologic abnormalities with continued use of the older term by some cytopathologists. Chart abstraction by a trained research assistant collected information regarding: (1) complicating medical conditions (immunosuppression, history of cancer, substance abuse, pregnancy); (2) history of cervical dysplasia; (3) primary care provider and number of visits to site; (4) the provider’s management plan; (5) all subsequent Pap test and colposcopy results; (6) evidence of communication of the management plan to the subject; and (7) demographics.

Accuracy of chart abstraction was confirmed by the lead author rescreening 20% of the records. The data were entered into a Paradox database (Borland Software Corporation, Scotts Valley, Calif), and statistical analysis was performed using SAS software (SAS Institute, Inc, Cary, NC) and Stata software (Stata Corporation, College Station, Tex).

Analysis

The goal of our analysis was to determine the factors associated with follow-up. We defined 3 mutually exclusive levels of adherence to the CPG: (1) complete adherence (colposcopy done if recommended or 3 repeat tests approximately every 6 months until 3 are normal or colposcopy if there is persistent abnormality); (2) moderate adherence (1 or 2 follow-up tests after the index Pap test); and (3) low adherence (no follow-up cytology or histology, or abnormality on repeat Pap test without colposcopy.)

Univariate analyses were performed relating the level of follow-up to the description of the abnormality (atypia vs ASCUS), patient demographics, and practice and provider characteristics (eg, training, availability of colposcopy on site, and plan of care). In analyses where cell size fell below 5, Fisher exact test or Yates correction (for analyses with row or column size greater than 2) replaced the Pearson chi-square. We performed multivariate analysis using a hierarchical model with patients nested within sites, and ordinal logistic regression to account for the 3 ordered levels of adherence to the CPG. Variables significant at P less than .2 in the univariate analyses were included in the multivariate model to ensure inclusion of potentially important variables. Also, insurance status was included, because others have found significant associations for this variable.^21-23 The provider’s management plan was not included as a variable in the regression model, because it is not sufficiently independent of the adherence levels.

Because subjects were clustered within 7 sites (not a simple random sample), we treated site as a random effect with patients nested within the site. An ordered logistic regression model was fit by maximum likelihood estimation, using the program GLLAMM⁶ in Stata version 6.0 The fit of the ordered logistic regression model to the data was tested using the method of Harrell and colleagues.²⁴ Briefly, this method compares observed mean values of the model’s linear combination of the predictor variables to expected values under the assumption that the model (and in particular the proportional odds ratio assumption) is correctly specified.

Results

of the 570 women identified with an atypical or ASCUS Pap test result in 1996, 554 charts were located and reviewed. The charts of 16 individuals could not be located for review. Some women (n=69) identified had a previous atypical or ASCUS Pap test result. The sample of women with no previous ASCUS or atypical result, excluding those with documented HIV infection (n=19) and those with a history of CIN or SIL (n=79), consisted of 387 subjects. The mean age of the subjects at the time of the index Pap test was 32 years (standard deviation=11.7). More than half of the subjects (58.4%) were Hispanic. The majority of subjects had Medicaid (73.1%). The majority of index Pap result abnormalities were described as atypia (61.5%), with the remainder ASCUS (38.5%). A small percentage of women (8.8%) were pregnant at the time of the index Pap test. The subjects were approximately evenly split between receiving care in medicine/pediatrics/obstetrics model sites and family medicine model sites. Colposcopy was provided on site (at least 2 times per week) in 4 of the centers that served 68.7% of the subjects.

Overall, many subjects remained in care in the health centers for a relatively short time. Almost one fourth of the subjects (23.7%) had a final visit to the health center less than 6 months from the index Pap test. Complete data including 2 full years of follow-up from the date of the index Pap test were available for 41% of subjects. Subjects with new atypia had a small chance of SIL on subsequent tests (6.5%, increasing to 10.9% for those subjects with persistent atypia). Results of biopsies for women with new and persistent atypia are presented in Table 1. The majority of subjects had normal or low-grade findings on histology, and no cancer was diagnosed.

Management of atypical results is depicted in the Figure 1. Providers were credited with a “correct” plan if they recommended either a repeat test in 6 months or less, or colposcopy. “Incorrect” management included any other recommendation or failure to document any plan (18.9%). Providers were most likely to recommend a repeat test in 3 months (31.8%) or 6 months (46.1%). We considered communication to subjects “complete” if the notes reflected discussion of the result at a visit (58.3%) or if a completed telephone call was documented (3.9%). Communication was considered attempted if a letter was sent (47.2%) or a call attempted (8.8%) and documented. In 8.3% of charts no documentation of attempted or completed communication was present. There were no significant associations of provider plan or communication of results with demographic variables, pregnancy, or abnormality type. Family physicians were more likely than midlevel providers or other physicians to generate a correct plan (87.3%, 80.8%, 69.8%, respectively; P=.002). Family physicians documented completed communication of results in 77.8% of cases, compared with 70.1% for other physicians and 57.4% for midlevel providers.

Providers were more likely to recommend colposcopy when the index Pap test result was reported as ASCUS rather than atypia (22.1% vs 11.4%, P=.016). Providers were much more likely to choose colposcopy if the pathologist made a recommendation for colposcopy rather than a repeat test (85.2% vs 8.9%, P=.001) and if the report included “favored SIL” in the result (75.0% vs 11.5%, P=.001.) When colposcopy was recommended, 70.8% completed the procedure. After a first atypical Pap test, 36% had no follow-up test; 27% had 1 follow-up test; 20% had 2 tests; and 17% had 3 subsequent tests. Subjects were found to have complete, moderate, and low adherence levels as follows: 26.9%, 27.6%, and 45.5%.

Univariate associations with levels of adherence are shown in Table 2. Complete adherence was more likely when the result was discussed at a visit, when the provider’s plan was documented, when the provider was other than a family physician, and if the plan was colposcopy as opposed to repeat cytology. Complete adherence was associated with remaining in care at the centers after 1997. It was more likely for results reported as ASCUS as opposed to atypia. Large differences are seen in the number of women with complete adherence by site, with the percentage ranging from 7% to 57%.

In the ordered logistic regression analysis, 4 factors were found to be predictive of the quality of adherence: the description of the abnormality (atypia vs ASCUS), availability of colposcopy on-site, discussion of the plan at a visit, and site. The detailed results are shown in Table 3. The adjusted odds ratio for complete versus moderate and low adherence (or equivalently moderate and complete vs low) associated with having an ASCUS lesion versus atypia is 1.56; for attending a clinic where colposcopy is referred off-site, 0.39; and for having discussed the follow-up plan, 3.44. The effects of ethnicity, type of health insurance, provider training, and pregnancy were not statistically significant.

Discussion

The introduction of the ASCUS label in the early 1990s created controversy and led to guidelines from the NCI7 and others^8,9 based primarily on expert opinion. Data from a large randomized clinical trial25 currently underway may resolve some of the questions about the effectiveness of repeat cytology versus immediate colposcopy in the management of low-grade dysplasia. Importantly, lack of evidence for the guideline may adversely affect compliance. Implementation of either strategy is difficult, especially in communities with many risk factors for poor follow-up. Our results indicate low adherence to the NCI interim guideline in an urban primary care setting serving predominately low-income minority women. Of concern is that almost half our sample, compared with 30% of subjects in a Canadian study,²⁶ had no follow-up cytology or histology. Our findings highlight some of the difficulties encountered in managing low-grade abnormalities in a setting challenged by high provider turnover, staffing difficulties, incorrect phone and address information, and financial and language barriers. Long-term repeated cytology is difficult to accomplish where loss to follow-up is common and competing demands of managing multiple chronic illnesses may interfere. Our findings are similar to those of other investigators^27,28 in that colposcopy, when recommended, was not completed by approximately one third of the women.

Our findings highlight the importance of strict use of Bethesda terminology by cytopathologists, since the older terminology may be misunderstood and less aggressively followed by primary care clinicians. We chose to include women with either atypical or ASCUS results, because they are different labels for the same spectrum of abnormalities. We found that providers were slightly more likely to recommend colposcopy for ASCUS and ultimately that women with atypical results were more likely to have low adherence to the CPG. This probably reflects that many primary care providers do not view the guideline as applicable to results reported as atypia or communicate with patients about atypical results differently.

There were substantial differences in rates of follow-up by site that are likely related to several factors. Although tracking of abnormal Pap test results was required at all sites the intensity varied, with the most successful site staffed by registered nurses who worked closely with the colposcopy providers in maintaining a manual card file. Sites where colposcopy was recommended more often had overall higher proportions of women with complete adherence to the CPG than those sites where serial cytology was the follow-up strategy of choice. Family physicians were more likely to generate a correct plan yet had worse overall adherence. This is likely due to a preference to follow up women by serial cytology, which proves difficult in settings where patients often stay in care for less than 2 years. However, the finding persists even after controlling for the provider plan, suggesting that other factors are also contributing to account for the differences between sites. Importantly, providers were more likely to opt for colposcopy, and women were more likely to complete it if the service was available in the subject’s health center.

We are unable to determine all of the causes of inadequate follow-up. Systemic factors, however, are clearly implicated by the present data, as indicated by the substantial variation by site and the number of charts in which a clear provider plan was not indicated or where no indication that the subject had been informed of the result was documented. When documentation was present that results were discussed with patients during visits patients were much more likely to have good follow-up, again suggesting that good communication of results is imperative. Our results also suggest that reducing barriers to colposcopy, by such means as providing the service on-site, may be effective in achieving optimal follow-up when colposcopy is recommended. Recommending colposcopy for all women with an initial ASCUS or atypical Pap test result will result in significant stress on available colposcopic resources and expose large numbers of women to an expensive, uncomfortable, and distressing procedure unnecessarily. However, providers need to carefully consider the risk of loss to follow-up in making a recommendation, take extra effort to insure that women are informed of and understand their Pap test result, and establish clinical tracking efforts to meet the challenge of serial cytology in high-risk settings.

Limitations

Our findings are limited in that the standard of care to which we compared practice is based on a guideline that is not evidence based. Other limitations of the project are largely due to constraints imposed by the abstraction of retrospective data from primary care charts. Information about how women were informed of results is sketchy, and failure to document may result in underestimation of actual patient contact. The extent to which women had follow-up if they no longer received care in the health centers cannot be estimated. Some women may have had follow-up with other providers outside our system. Demographic information was obtained from site registration databases rather than self-report and may include some misclassification. Some important information is not captured consistently in the site’s computers, including changes in insurance and primary care providers. Provider turnover was extremely high in these health centers during the interval studied, though lists of providers are not available to create a precise measure to include in our model.

Acknowledgments

This work was supported by a clinical research training grant from the American Cancer Society.

Approximately 50 million Papanicolaou (Pap) tests are performed annually in the United States, of which approximately 5% will require further evaluation because of an abnormality.¹ The Bethesda System for reporting cervical/vaginal cytologic diagnoses,² developed in 1988 and revised in 1991, provides uniform diagnostic terminology to improve consistency of reporting. The Bethesda System introduced new terminology for atypical cytology, creating the label “atypical squamous cells of uncertain significance” (ASCUS). Pap test results labelled ASCUS encompass a spectrum of cellular change reflecting a variety of pathologic processes that cannot be more specifically categorized, including reactive changes, inflammation, human papillomavirus (HPV) related changes, suggested dysplasia, and less than optimal slide preparation.² Because studies have consistently shown a small but real risk that an ASCUS Pap result represents an underlying high-grade lesion,^3-6 there has been much debate about their management. Uncertainty about optimal management led to the development of clinical practice guidelines (CPG).^7-9

The National Cancer Institute⁷ (NCI) and a Canadian expert pane¹⁸ suggest that ASCUS and low-grade squamous intraepithelial lesions can be managed either by immediate colposcopy or by repeating cervical cytology. Clinicians may, if they choose, reserve referral to colposcopy for those patients with persistent abnormality on subsequent cytology. This conservative strategy is based on several assumptions:¹⁰ (1) conditions that mimic dysplasia, including reactive changes, are common; (2) spontaneous regression occurs in at least 50% of low-grade lesions; (3) the rate of progression is relatively slow; (4) the chance of carcinoma in situ with an ASCUS Pap test result is low; and (5) persistently negative follow-up Pap results are unlikely to be false negatives. The current practice guideline suggests that if clinicians opt for cytologic follow-up of ASCUS, repeat Pap tests should be done every 4 to 6 months until 3 consecutive normal test results have been obtained. Colposcopy should be performed if there is a subsequent abnormal Pap test result.⁷ The 1996 statement of the American College of Obstetricians and Gynecologists is similar, with the addition that colposcopy be performed for women with a single ASCUS Pap result if they also have any of the following high-risk factors: HPV infection, human immunodeficiency virus (HIV) infection, smoking, or multiple sexual partners.⁹

A disproportionate share of the cervical cancer burden in this country, in terms of both incidence and mortality, is shouldered by low-income minority women.^11-13 Minority and low-income populations have the highest rates of nonadherence with repeat Pap test or colposcopy.^14-18 Thus, the women who are at the highest risk of cervical intraepithelial neoplasia (CIN) may also be the most at risk for suboptimal management after an abnormal Pap test result is obtained. This project assessed current management of ASCUS Pap results in a diverse community at high risk of cervical cancer by comparing actual practice to the NCI guideline. In addition to assessing levels of follow-up in this community, we sought to identify predictors of adherence to that guideline.

Methods

Setting and Population

Data were collected in 7 community health centers serving a low-income, predominantly minority population. The centers represent diversity with regard to practice type (medicine/pediatrics/obstetrics model or family medicine model) and size (20,000-80,000 visits/year). The subjects were women with newly identified atypia or ASCUS Pap test results in 1996; women for whom the index Pap test represented persistent atypia or ASCUS were excluded, as were women with a history of CIN or squamous intraepithelial lesion (SIL). Those who were known to be infected with HIV represent a special category^19,20 that may have been managed differently at the clinician’s discretion; therefore, they were also excluded from the analysis.

Data Collection

After approval by the appropriate institutional review boards, women with atypical results in 1996 were identified by 2 mechanisms: site logs maintained for clinical tracking, and because all sites exclusively use the same laboratory, the central cytology database. Pap test results described as atypia or ASCUS were included, as these represent different terminology for the same spectrum of cytologic abnormalities with continued use of the older term by some cytopathologists. Chart abstraction by a trained research assistant collected information regarding: (1) complicating medical conditions (immunosuppression, history of cancer, substance abuse, pregnancy); (2) history of cervical dysplasia; (3) primary care provider and number of visits to site; (4) the provider’s management plan; (5) all subsequent Pap test and colposcopy results; (6) evidence of communication of the management plan to the subject; and (7) demographics.

Accuracy of chart abstraction was confirmed by the lead author rescreening 20% of the records. The data were entered into a Paradox database (Borland Software Corporation, Scotts Valley, Calif), and statistical analysis was performed using SAS software (SAS Institute, Inc, Cary, NC) and Stata software (Stata Corporation, College Station, Tex).

Analysis

The goal of our analysis was to determine the factors associated with follow-up. We defined 3 mutually exclusive levels of adherence to the CPG: (1) complete adherence (colposcopy done if recommended or 3 repeat tests approximately every 6 months until 3 are normal or colposcopy if there is persistent abnormality); (2) moderate adherence (1 or 2 follow-up tests after the index Pap test); and (3) low adherence (no follow-up cytology or histology, or abnormality on repeat Pap test without colposcopy.)

Univariate analyses were performed relating the level of follow-up to the description of the abnormality (atypia vs ASCUS), patient demographics, and practice and provider characteristics (eg, training, availability of colposcopy on site, and plan of care). In analyses where cell size fell below 5, Fisher exact test or Yates correction (for analyses with row or column size greater than 2) replaced the Pearson chi-square. We performed multivariate analysis using a hierarchical model with patients nested within sites, and ordinal logistic regression to account for the 3 ordered levels of adherence to the CPG. Variables significant at P less than .2 in the univariate analyses were included in the multivariate model to ensure inclusion of potentially important variables. Also, insurance status was included, because others have found significant associations for this variable.^21-23 The provider’s management plan was not included as a variable in the regression model, because it is not sufficiently independent of the adherence levels.

Because subjects were clustered within 7 sites (not a simple random sample), we treated site as a random effect with patients nested within the site. An ordered logistic regression model was fit by maximum likelihood estimation, using the program GLLAMM⁶ in Stata version 6.0 The fit of the ordered logistic regression model to the data was tested using the method of Harrell and colleagues.²⁴ Briefly, this method compares observed mean values of the model’s linear combination of the predictor variables to expected values under the assumption that the model (and in particular the proportional odds ratio assumption) is correctly specified.

Results

of the 570 women identified with an atypical or ASCUS Pap test result in 1996, 554 charts were located and reviewed. The charts of 16 individuals could not be located for review. Some women (n=69) identified had a previous atypical or ASCUS Pap test result. The sample of women with no previous ASCUS or atypical result, excluding those with documented HIV infection (n=19) and those with a history of CIN or SIL (n=79), consisted of 387 subjects. The mean age of the subjects at the time of the index Pap test was 32 years (standard deviation=11.7). More than half of the subjects (58.4%) were Hispanic. The majority of subjects had Medicaid (73.1%). The majority of index Pap result abnormalities were described as atypia (61.5%), with the remainder ASCUS (38.5%). A small percentage of women (8.8%) were pregnant at the time of the index Pap test. The subjects were approximately evenly split between receiving care in medicine/pediatrics/obstetrics model sites and family medicine model sites. Colposcopy was provided on site (at least 2 times per week) in 4 of the centers that served 68.7% of the subjects.

Overall, many subjects remained in care in the health centers for a relatively short time. Almost one fourth of the subjects (23.7%) had a final visit to the health center less than 6 months from the index Pap test. Complete data including 2 full years of follow-up from the date of the index Pap test were available for 41% of subjects. Subjects with new atypia had a small chance of SIL on subsequent tests (6.5%, increasing to 10.9% for those subjects with persistent atypia). Results of biopsies for women with new and persistent atypia are presented in Table 1. The majority of subjects had normal or low-grade findings on histology, and no cancer was diagnosed.

Management of atypical results is depicted in the Figure 1. Providers were credited with a “correct” plan if they recommended either a repeat test in 6 months or less, or colposcopy. “Incorrect” management included any other recommendation or failure to document any plan (18.9%). Providers were most likely to recommend a repeat test in 3 months (31.8%) or 6 months (46.1%). We considered communication to subjects “complete” if the notes reflected discussion of the result at a visit (58.3%) or if a completed telephone call was documented (3.9%). Communication was considered attempted if a letter was sent (47.2%) or a call attempted (8.8%) and documented. In 8.3% of charts no documentation of attempted or completed communication was present. There were no significant associations of provider plan or communication of results with demographic variables, pregnancy, or abnormality type. Family physicians were more likely than midlevel providers or other physicians to generate a correct plan (87.3%, 80.8%, 69.8%, respectively; P=.002). Family physicians documented completed communication of results in 77.8% of cases, compared with 70.1% for other physicians and 57.4% for midlevel providers.

Providers were more likely to recommend colposcopy when the index Pap test result was reported as ASCUS rather than atypia (22.1% vs 11.4%, P=.016). Providers were much more likely to choose colposcopy if the pathologist made a recommendation for colposcopy rather than a repeat test (85.2% vs 8.9%, P=.001) and if the report included “favored SIL” in the result (75.0% vs 11.5%, P=.001.) When colposcopy was recommended, 70.8% completed the procedure. After a first atypical Pap test, 36% had no follow-up test; 27% had 1 follow-up test; 20% had 2 tests; and 17% had 3 subsequent tests. Subjects were found to have complete, moderate, and low adherence levels as follows: 26.9%, 27.6%, and 45.5%.

Univariate associations with levels of adherence are shown in Table 2. Complete adherence was more likely when the result was discussed at a visit, when the provider’s plan was documented, when the provider was other than a family physician, and if the plan was colposcopy as opposed to repeat cytology. Complete adherence was associated with remaining in care at the centers after 1997. It was more likely for results reported as ASCUS as opposed to atypia. Large differences are seen in the number of women with complete adherence by site, with the percentage ranging from 7% to 57%.

In the ordered logistic regression analysis, 4 factors were found to be predictive of the quality of adherence: the description of the abnormality (atypia vs ASCUS), availability of colposcopy on-site, discussion of the plan at a visit, and site. The detailed results are shown in Table 3. The adjusted odds ratio for complete versus moderate and low adherence (or equivalently moderate and complete vs low) associated with having an ASCUS lesion versus atypia is 1.56; for attending a clinic where colposcopy is referred off-site, 0.39; and for having discussed the follow-up plan, 3.44. The effects of ethnicity, type of health insurance, provider training, and pregnancy were not statistically significant.

Discussion

The introduction of the ASCUS label in the early 1990s created controversy and led to guidelines from the NCI7 and others^8,9 based primarily on expert opinion. Data from a large randomized clinical trial25 currently underway may resolve some of the questions about the effectiveness of repeat cytology versus immediate colposcopy in the management of low-grade dysplasia. Importantly, lack of evidence for the guideline may adversely affect compliance. Implementation of either strategy is difficult, especially in communities with many risk factors for poor follow-up. Our results indicate low adherence to the NCI interim guideline in an urban primary care setting serving predominately low-income minority women. Of concern is that almost half our sample, compared with 30% of subjects in a Canadian study,²⁶ had no follow-up cytology or histology. Our findings highlight some of the difficulties encountered in managing low-grade abnormalities in a setting challenged by high provider turnover, staffing difficulties, incorrect phone and address information, and financial and language barriers. Long-term repeated cytology is difficult to accomplish where loss to follow-up is common and competing demands of managing multiple chronic illnesses may interfere. Our findings are similar to those of other investigators^27,28 in that colposcopy, when recommended, was not completed by approximately one third of the women.

Our findings highlight the importance of strict use of Bethesda terminology by cytopathologists, since the older terminology may be misunderstood and less aggressively followed by primary care clinicians. We chose to include women with either atypical or ASCUS results, because they are different labels for the same spectrum of abnormalities. We found that providers were slightly more likely to recommend colposcopy for ASCUS and ultimately that women with atypical results were more likely to have low adherence to the CPG. This probably reflects that many primary care providers do not view the guideline as applicable to results reported as atypia or communicate with patients about atypical results differently.

There were substantial differences in rates of follow-up by site that are likely related to several factors. Although tracking of abnormal Pap test results was required at all sites the intensity varied, with the most successful site staffed by registered nurses who worked closely with the colposcopy providers in maintaining a manual card file. Sites where colposcopy was recommended more often had overall higher proportions of women with complete adherence to the CPG than those sites where serial cytology was the follow-up strategy of choice. Family physicians were more likely to generate a correct plan yet had worse overall adherence. This is likely due to a preference to follow up women by serial cytology, which proves difficult in settings where patients often stay in care for less than 2 years. However, the finding persists even after controlling for the provider plan, suggesting that other factors are also contributing to account for the differences between sites. Importantly, providers were more likely to opt for colposcopy, and women were more likely to complete it if the service was available in the subject’s health center.

We are unable to determine all of the causes of inadequate follow-up. Systemic factors, however, are clearly implicated by the present data, as indicated by the substantial variation by site and the number of charts in which a clear provider plan was not indicated or where no indication that the subject had been informed of the result was documented. When documentation was present that results were discussed with patients during visits patients were much more likely to have good follow-up, again suggesting that good communication of results is imperative. Our results also suggest that reducing barriers to colposcopy, by such means as providing the service on-site, may be effective in achieving optimal follow-up when colposcopy is recommended. Recommending colposcopy for all women with an initial ASCUS or atypical Pap test result will result in significant stress on available colposcopic resources and expose large numbers of women to an expensive, uncomfortable, and distressing procedure unnecessarily. However, providers need to carefully consider the risk of loss to follow-up in making a recommendation, take extra effort to insure that women are informed of and understand their Pap test result, and establish clinical tracking efforts to meet the challenge of serial cytology in high-risk settings.

Limitations

Our findings are limited in that the standard of care to which we compared practice is based on a guideline that is not evidence based. Other limitations of the project are largely due to constraints imposed by the abstraction of retrospective data from primary care charts. Information about how women were informed of results is sketchy, and failure to document may result in underestimation of actual patient contact. The extent to which women had follow-up if they no longer received care in the health centers cannot be estimated. Some women may have had follow-up with other providers outside our system. Demographic information was obtained from site registration databases rather than self-report and may include some misclassification. Some important information is not captured consistently in the site’s computers, including changes in insurance and primary care providers. Provider turnover was extremely high in these health centers during the interval studied, though lists of providers are not available to create a precise measure to include in our model.

Acknowledgments

This work was supported by a clinical research training grant from the American Cancer Society.

Much has been learned about how to effectively treat depressive disorders, but we remain less certain about how to deliver these treatments in routine primary care practice. Clinical guidelines point the way; their implementation, however, requires system change. Key issues in improving health care system capacities for effective depression care include: (1) enhancing continuity of care; (2) activating and empowering patients; (3) matching interventions to patients, including stepped care strategies; (4) improving treatment follow through; (5) monitoring clinical course outcomes; and (6) revising the structure of care so guideline-based care is feasible in routine patient care. (J Fam Pract 2001; 50:535-537)

The public health significance of depression as a prevalent disorder in primary care practice is well established, and the value of diagnosing this disorder accurately and treating it effectively is amply documented.¹ However, the benefits of treatment that are evident in research projects remain elusive in routine ambulatory practice. The outcomes of usual care are still significantly worse than those of standardized interventions carefully monitored by investigators.² Even the benefits of the latter dissipate when the research program is completed.³ The critical issue, therefore, is how to transfer efficacious treatments of depression from research settings to routine primary care practice in a manner that will permit them to flourish. Recommendations about steps pertinent to these processes are presented in this paper.

Guideline-Based Treatment Algorithms

The treatment algorithms recommended in 1993⁴ and 2000⁵ by the American Psychiatric Association and in 1993 by the Depression Guideline Panel of the Agency for Health Care Policy and Research (AHCPR)¹ are landmarks in the quest for optimal management of depressive disorders. Despite initial concerns about the effectiveness of treatments transferred from the psychiatric to the primary care sector, the guideline recommendations have proven valid and durable.⁶ It is clear that antidepressant medications produce a 60% recovery rate when prescribed within proper dosages and for adequate duration. Depression-specific time-limited psychotherapies achieve similar outcomes, even with patients experiencing moderate to severe symptomatology.⁷ Two principles emerge from this body of work: (1) major depression should not be treated with anxiolytic medications alone or with long-term psychotherapy; and (2) patient preference for a particular guideline-based treatment should be considered when it is clinically and practically feasible.

Despite this scientific progress and the extensive efforts to disseminate the AHCPR depression guidelines, few would assert that such efforts have significantly influenced routine primary care of depression. Thus, guidelines are a necessary but not sufficient condition for improving the treatment of depression; they constitute a blueprint for building rather than completing a structure. Accordingly, attention has shifted to the manner in which general principles of effective clinical care can be customized to the structural, fiscal, and sociodemographic characteristics of a particular primary care practice. The study by Wells and colleagues⁸ illustrates efforts to disseminate guidelines in the context of local circumstances and suggests that administrative support for state of the art practice can positively influence patient outcomes.

Moving Beyond the Guidelines

The dissemination of existing guidelines surely is warranted, but it should be recognized that guideline standard treatments are imperfect. Antidepressant medications and depression-specific psychotherapies produce only 70% to 75% recovery rates even in treatment-completer analyses.¹ Thus, altered strategies, like those that follow, are needed to help the significant minority of depressed patients who fail to recover.

Continunity of Care

Although case identification attracted much attention during the 1980s, various studies have demonstrated that an improved case finding by itself does not improve depression outcomes.^7,8 Inadequate attention has been directed to the fuller continuum of care on which screening and assessment are the starting points. Case identification must be followed by timely and targeted feedback to the primary care physician; appropriate treatment must be provided the patient; the patient’s clinical course must be actively monitored; and the treatment should be modified when clinically indicated. Each component of this continuum must be refined if the quality of care is to improve.

Activating and Empowering Patients

It is vital that patients with depression be actively involved in their treatment, since overcoming helplessness and hopelessness is central to recovery.¹ Strategies for activating and empowering these patients range from the educational to the social. The former include increasingly effective materials about the nature and course of mood disorders that use state of the art technologies to capture and maintain the patient’s interest. Self-care programs have been developed that permit patients to monitor beliefs and behaviors linked to clinical depression. Social strategies for activating and empowering patients are also becoming more commonplace.

Matching Patient to Intervention

Uncomplicated cases of depression are effectively treated with presently available interventions, but we are less successful in treating patients whose particular clinical or demographic characteristics make their depression atypical. The resulting uncertainties have stimulated clinical trials of interventions for co-existing Axis I and II psychopathology, double depression, minor depression and dysthymia, and mood disorders that co-exist with physical illnesses. Treatments compatible with the specific needs and expectations of racial or ethnic minorities are also the focus of clinical trials.¹¹

Ambiguities also persist as to which patient subgroups require particular treatment decisions. For which patients is “watchful waiting” the appropriate treatment, and which patients require immediate referral to a mental health specialist? Also, patient subgroups requiring customized interventions are made up of high users of medical care¹² and nonresponders to initial treatments. Collaborative care in which primary care physicians and mental health specialists co-manage patients^13,14 and stepped care in which the intensity and content of treatment are guided by initial outcomes¹⁵ are emerging as effective approaches to depression care. Of particular interest with regard to stepped care is the relationship between the intensity of treatment and clinical outcome (ie, the marginal point at which optimal cost-effectiveness is achieved).

Treatment Follow-Through

Efficacious treatments are necessary to improve the care of depression, but it is equally vital that patients follow through with these treatments. Efforts such as psychoeducation, motivational strategies and family involvement may improve fuller patient participation. However, we still do not adequately understand the influence of such variables as socioeconomic status, race, and ethnicity on treatment participation. More specifically, do patients with particular characteristics enter treatment with implicit or explicit expectations at variance with the actual treatment process? Studies of the congruence between a patient’s illness model and the treatment offered by the physician are recommended.

Monitoring Clinical Course and Outcome

A depressive episode’s acute phase typically attracts much clinical scrutiny from primary care physicians and other providers. However, this scrutiny must extend well beyond the initial 6 to 12 weeks of acute-phase treatment, since mood disorders are often chronic in nature. As this growing awareness extends the duration for monitoring mood disorders, various clinical and practical decisions will be required since all too scarce resources are needed to longitudinally assess a patient’s depressed state and level of functioning. A specific schedule is needed, given that the time frame for improved social functioning exceeds that for symptom resolution. Also, decisions are required with regard to how the information needed for monitoring purposes will be obtained. In-person assessments are preferable but impractical when patients live at a distance from the health center, have limited mobility, or find it inconvenient to miss work. It is of interest, therefore, that relatively efficient telephone follow-up assessments yield reliable and valuable information.^16,17

Revising the Structure of Care

Although refined treatments can improve patient outcomes, more fundamental change in the structure of ambulatory medicine is needed if desired outcomes are to be regularly achieved.¹⁸ This perspective questions the present structure of primary care as it pertains to treating depression. For example, given the constraints imposed by the typical brief 10- to 15-minute physician-patient encounter, is the leadership of primary care willing to sanction structural changes more conducive to the treatment of depression? Can the leadership judge proper management of mood disorder an opportunity to improve the patient’s health rather than an additional task imposed on already overburdened physicians?

A central element in the needed structural change is the creation of fiscal and cultural incentives to match such disincentives as capitated coverage and mental health carve-outs. A second structural element potentially amenable to change is the role of nurses and social workers already functioning within primary care settings. Can they expand their job description to include responsibility for depressive disorders, as well as such conditions as hypertension, diabetes, and asthma, and perhaps even come to serve as chronic disease specialists?¹⁹ Various studies are analyzing the tasks to be performed routinely by nurses, such as monitoring a patient’s adherence to prescribed antidepressant medication regimens; assessing patients’ clinical status, providing feedback to primary care providers, ensuring that guideline-based services are provided, and conveying interest and concern to depressed patients who may feel helpless and even hopeless.²⁰

Conclusions

Modified treatment algorithms can enhance the effectiveness of available interventions, but questions about their compatibility with the present organizational structure of primary care practice ultimately must be confronted and resolved. The manner in which this occurs will determine whether research findings from the laboratory can be transferred to the ambulatory medical setting in ways that benefit the large numbers of depressed patients in primary care settings.

Much has been learned about how to effectively treat depressive disorders, but we remain less certain about how to deliver these treatments in routine primary care practice. Clinical guidelines point the way; their implementation, however, requires system change. Key issues in improving health care system capacities for effective depression care include: (1) enhancing continuity of care; (2) activating and empowering patients; (3) matching interventions to patients, including stepped care strategies; (4) improving treatment follow through; (5) monitoring clinical course outcomes; and (6) revising the structure of care so guideline-based care is feasible in routine patient care. (J Fam Pract 2001; 50:535-537)

The public health significance of depression as a prevalent disorder in primary care practice is well established, and the value of diagnosing this disorder accurately and treating it effectively is amply documented.¹ However, the benefits of treatment that are evident in research projects remain elusive in routine ambulatory practice. The outcomes of usual care are still significantly worse than those of standardized interventions carefully monitored by investigators.² Even the benefits of the latter dissipate when the research program is completed.³ The critical issue, therefore, is how to transfer efficacious treatments of depression from research settings to routine primary care practice in a manner that will permit them to flourish. Recommendations about steps pertinent to these processes are presented in this paper.

Guideline-Based Treatment Algorithms

The treatment algorithms recommended in 1993⁴ and 2000⁵ by the American Psychiatric Association and in 1993 by the Depression Guideline Panel of the Agency for Health Care Policy and Research (AHCPR)¹ are landmarks in the quest for optimal management of depressive disorders. Despite initial concerns about the effectiveness of treatments transferred from the psychiatric to the primary care sector, the guideline recommendations have proven valid and durable.⁶ It is clear that antidepressant medications produce a 60% recovery rate when prescribed within proper dosages and for adequate duration. Depression-specific time-limited psychotherapies achieve similar outcomes, even with patients experiencing moderate to severe symptomatology.⁷ Two principles emerge from this body of work: (1) major depression should not be treated with anxiolytic medications alone or with long-term psychotherapy; and (2) patient preference for a particular guideline-based treatment should be considered when it is clinically and practically feasible.

Despite this scientific progress and the extensive efforts to disseminate the AHCPR depression guidelines, few would assert that such efforts have significantly influenced routine primary care of depression. Thus, guidelines are a necessary but not sufficient condition for improving the treatment of depression; they constitute a blueprint for building rather than completing a structure. Accordingly, attention has shifted to the manner in which general principles of effective clinical care can be customized to the structural, fiscal, and sociodemographic characteristics of a particular primary care practice. The study by Wells and colleagues⁸ illustrates efforts to disseminate guidelines in the context of local circumstances and suggests that administrative support for state of the art practice can positively influence patient outcomes.

Moving Beyond the Guidelines

The dissemination of existing guidelines surely is warranted, but it should be recognized that guideline standard treatments are imperfect. Antidepressant medications and depression-specific psychotherapies produce only 70% to 75% recovery rates even in treatment-completer analyses.¹ Thus, altered strategies, like those that follow, are needed to help the significant minority of depressed patients who fail to recover.

Continunity of Care

Although case identification attracted much attention during the 1980s, various studies have demonstrated that an improved case finding by itself does not improve depression outcomes.^7,8 Inadequate attention has been directed to the fuller continuum of care on which screening and assessment are the starting points. Case identification must be followed by timely and targeted feedback to the primary care physician; appropriate treatment must be provided the patient; the patient’s clinical course must be actively monitored; and the treatment should be modified when clinically indicated. Each component of this continuum must be refined if the quality of care is to improve.

Activating and Empowering Patients

It is vital that patients with depression be actively involved in their treatment, since overcoming helplessness and hopelessness is central to recovery.¹ Strategies for activating and empowering these patients range from the educational to the social. The former include increasingly effective materials about the nature and course of mood disorders that use state of the art technologies to capture and maintain the patient’s interest. Self-care programs have been developed that permit patients to monitor beliefs and behaviors linked to clinical depression. Social strategies for activating and empowering patients are also becoming more commonplace.

Matching Patient to Intervention

Uncomplicated cases of depression are effectively treated with presently available interventions, but we are less successful in treating patients whose particular clinical or demographic characteristics make their depression atypical. The resulting uncertainties have stimulated clinical trials of interventions for co-existing Axis I and II psychopathology, double depression, minor depression and dysthymia, and mood disorders that co-exist with physical illnesses. Treatments compatible with the specific needs and expectations of racial or ethnic minorities are also the focus of clinical trials.¹¹

Ambiguities also persist as to which patient subgroups require particular treatment decisions. For which patients is “watchful waiting” the appropriate treatment, and which patients require immediate referral to a mental health specialist? Also, patient subgroups requiring customized interventions are made up of high users of medical care¹² and nonresponders to initial treatments. Collaborative care in which primary care physicians and mental health specialists co-manage patients^13,14 and stepped care in which the intensity and content of treatment are guided by initial outcomes¹⁵ are emerging as effective approaches to depression care. Of particular interest with regard to stepped care is the relationship between the intensity of treatment and clinical outcome (ie, the marginal point at which optimal cost-effectiveness is achieved).

Treatment Follow-Through

Efficacious treatments are necessary to improve the care of depression, but it is equally vital that patients follow through with these treatments. Efforts such as psychoeducation, motivational strategies and family involvement may improve fuller patient participation. However, we still do not adequately understand the influence of such variables as socioeconomic status, race, and ethnicity on treatment participation. More specifically, do patients with particular characteristics enter treatment with implicit or explicit expectations at variance with the actual treatment process? Studies of the congruence between a patient’s illness model and the treatment offered by the physician are recommended.

Monitoring Clinical Course and Outcome

A depressive episode’s acute phase typically attracts much clinical scrutiny from primary care physicians and other providers. However, this scrutiny must extend well beyond the initial 6 to 12 weeks of acute-phase treatment, since mood disorders are often chronic in nature. As this growing awareness extends the duration for monitoring mood disorders, various clinical and practical decisions will be required since all too scarce resources are needed to longitudinally assess a patient’s depressed state and level of functioning. A specific schedule is needed, given that the time frame for improved social functioning exceeds that for symptom resolution. Also, decisions are required with regard to how the information needed for monitoring purposes will be obtained. In-person assessments are preferable but impractical when patients live at a distance from the health center, have limited mobility, or find it inconvenient to miss work. It is of interest, therefore, that relatively efficient telephone follow-up assessments yield reliable and valuable information.^16,17

Revising the Structure of Care

Although refined treatments can improve patient outcomes, more fundamental change in the structure of ambulatory medicine is needed if desired outcomes are to be regularly achieved.¹⁸ This perspective questions the present structure of primary care as it pertains to treating depression. For example, given the constraints imposed by the typical brief 10- to 15-minute physician-patient encounter, is the leadership of primary care willing to sanction structural changes more conducive to the treatment of depression? Can the leadership judge proper management of mood disorder an opportunity to improve the patient’s health rather than an additional task imposed on already overburdened physicians?

A central element in the needed structural change is the creation of fiscal and cultural incentives to match such disincentives as capitated coverage and mental health carve-outs. A second structural element potentially amenable to change is the role of nurses and social workers already functioning within primary care settings. Can they expand their job description to include responsibility for depressive disorders, as well as such conditions as hypertension, diabetes, and asthma, and perhaps even come to serve as chronic disease specialists?¹⁹ Various studies are analyzing the tasks to be performed routinely by nurses, such as monitoring a patient’s adherence to prescribed antidepressant medication regimens; assessing patients’ clinical status, providing feedback to primary care providers, ensuring that guideline-based services are provided, and conveying interest and concern to depressed patients who may feel helpless and even hopeless.²⁰

Conclusions

Modified treatment algorithms can enhance the effectiveness of available interventions, but questions about their compatibility with the present organizational structure of primary care practice ultimately must be confronted and resolved. The manner in which this occurs will determine whether research findings from the laboratory can be transferred to the ambulatory medical setting in ways that benefit the large numbers of depressed patients in primary care settings.

Much has been learned about how to effectively treat depressive disorders, but we remain less certain about how to deliver these treatments in routine primary care practice. Clinical guidelines point the way; their implementation, however, requires system change. Key issues in improving health care system capacities for effective depression care include: (1) enhancing continuity of care; (2) activating and empowering patients; (3) matching interventions to patients, including stepped care strategies; (4) improving treatment follow through; (5) monitoring clinical course outcomes; and (6) revising the structure of care so guideline-based care is feasible in routine patient care. (J Fam Pract 2001; 50:535-537)

The public health significance of depression as a prevalent disorder in primary care practice is well established, and the value of diagnosing this disorder accurately and treating it effectively is amply documented.¹ However, the benefits of treatment that are evident in research projects remain elusive in routine ambulatory practice. The outcomes of usual care are still significantly worse than those of standardized interventions carefully monitored by investigators.² Even the benefits of the latter dissipate when the research program is completed.³ The critical issue, therefore, is how to transfer efficacious treatments of depression from research settings to routine primary care practice in a manner that will permit them to flourish. Recommendations about steps pertinent to these processes are presented in this paper.

Guideline-Based Treatment Algorithms

The treatment algorithms recommended in 1993⁴ and 2000⁵ by the American Psychiatric Association and in 1993 by the Depression Guideline Panel of the Agency for Health Care Policy and Research (AHCPR)¹ are landmarks in the quest for optimal management of depressive disorders. Despite initial concerns about the effectiveness of treatments transferred from the psychiatric to the primary care sector, the guideline recommendations have proven valid and durable.⁶ It is clear that antidepressant medications produce a 60% recovery rate when prescribed within proper dosages and for adequate duration. Depression-specific time-limited psychotherapies achieve similar outcomes, even with patients experiencing moderate to severe symptomatology.⁷ Two principles emerge from this body of work: (1) major depression should not be treated with anxiolytic medications alone or with long-term psychotherapy; and (2) patient preference for a particular guideline-based treatment should be considered when it is clinically and practically feasible.

Despite this scientific progress and the extensive efforts to disseminate the AHCPR depression guidelines, few would assert that such efforts have significantly influenced routine primary care of depression. Thus, guidelines are a necessary but not sufficient condition for improving the treatment of depression; they constitute a blueprint for building rather than completing a structure. Accordingly, attention has shifted to the manner in which general principles of effective clinical care can be customized to the structural, fiscal, and sociodemographic characteristics of a particular primary care practice. The study by Wells and colleagues⁸ illustrates efforts to disseminate guidelines in the context of local circumstances and suggests that administrative support for state of the art practice can positively influence patient outcomes.

Moving Beyond the Guidelines

The dissemination of existing guidelines surely is warranted, but it should be recognized that guideline standard treatments are imperfect. Antidepressant medications and depression-specific psychotherapies produce only 70% to 75% recovery rates even in treatment-completer analyses.¹ Thus, altered strategies, like those that follow, are needed to help the significant minority of depressed patients who fail to recover.

Continunity of Care

Although case identification attracted much attention during the 1980s, various studies have demonstrated that an improved case finding by itself does not improve depression outcomes.^7,8 Inadequate attention has been directed to the fuller continuum of care on which screening and assessment are the starting points. Case identification must be followed by timely and targeted feedback to the primary care physician; appropriate treatment must be provided the patient; the patient’s clinical course must be actively monitored; and the treatment should be modified when clinically indicated. Each component of this continuum must be refined if the quality of care is to improve.

Activating and Empowering Patients

It is vital that patients with depression be actively involved in their treatment, since overcoming helplessness and hopelessness is central to recovery.¹ Strategies for activating and empowering these patients range from the educational to the social. The former include increasingly effective materials about the nature and course of mood disorders that use state of the art technologies to capture and maintain the patient’s interest. Self-care programs have been developed that permit patients to monitor beliefs and behaviors linked to clinical depression. Social strategies for activating and empowering patients are also becoming more commonplace.

Matching Patient to Intervention

Uncomplicated cases of depression are effectively treated with presently available interventions, but we are less successful in treating patients whose particular clinical or demographic characteristics make their depression atypical. The resulting uncertainties have stimulated clinical trials of interventions for co-existing Axis I and II psychopathology, double depression, minor depression and dysthymia, and mood disorders that co-exist with physical illnesses. Treatments compatible with the specific needs and expectations of racial or ethnic minorities are also the focus of clinical trials.¹¹

Ambiguities also persist as to which patient subgroups require particular treatment decisions. For which patients is “watchful waiting” the appropriate treatment, and which patients require immediate referral to a mental health specialist? Also, patient subgroups requiring customized interventions are made up of high users of medical care¹² and nonresponders to initial treatments. Collaborative care in which primary care physicians and mental health specialists co-manage patients^13,14 and stepped care in which the intensity and content of treatment are guided by initial outcomes¹⁵ are emerging as effective approaches to depression care. Of particular interest with regard to stepped care is the relationship between the intensity of treatment and clinical outcome (ie, the marginal point at which optimal cost-effectiveness is achieved).

Treatment Follow-Through

Efficacious treatments are necessary to improve the care of depression, but it is equally vital that patients follow through with these treatments. Efforts such as psychoeducation, motivational strategies and family involvement may improve fuller patient participation. However, we still do not adequately understand the influence of such variables as socioeconomic status, race, and ethnicity on treatment participation. More specifically, do patients with particular characteristics enter treatment with implicit or explicit expectations at variance with the actual treatment process? Studies of the congruence between a patient’s illness model and the treatment offered by the physician are recommended.

Monitoring Clinical Course and Outcome

A depressive episode’s acute phase typically attracts much clinical scrutiny from primary care physicians and other providers. However, this scrutiny must extend well beyond the initial 6 to 12 weeks of acute-phase treatment, since mood disorders are often chronic in nature. As this growing awareness extends the duration for monitoring mood disorders, various clinical and practical decisions will be required since all too scarce resources are needed to longitudinally assess a patient’s depressed state and level of functioning. A specific schedule is needed, given that the time frame for improved social functioning exceeds that for symptom resolution. Also, decisions are required with regard to how the information needed for monitoring purposes will be obtained. In-person assessments are preferable but impractical when patients live at a distance from the health center, have limited mobility, or find it inconvenient to miss work. It is of interest, therefore, that relatively efficient telephone follow-up assessments yield reliable and valuable information.^16,17

Revising the Structure of Care

Although refined treatments can improve patient outcomes, more fundamental change in the structure of ambulatory medicine is needed if desired outcomes are to be regularly achieved.¹⁸ This perspective questions the present structure of primary care as it pertains to treating depression. For example, given the constraints imposed by the typical brief 10- to 15-minute physician-patient encounter, is the leadership of primary care willing to sanction structural changes more conducive to the treatment of depression? Can the leadership judge proper management of mood disorder an opportunity to improve the patient’s health rather than an additional task imposed on already overburdened physicians?

A central element in the needed structural change is the creation of fiscal and cultural incentives to match such disincentives as capitated coverage and mental health carve-outs. A second structural element potentially amenable to change is the role of nurses and social workers already functioning within primary care settings. Can they expand their job description to include responsibility for depressive disorders, as well as such conditions as hypertension, diabetes, and asthma, and perhaps even come to serve as chronic disease specialists?¹⁹ Various studies are analyzing the tasks to be performed routinely by nurses, such as monitoring a patient’s adherence to prescribed antidepressant medication regimens; assessing patients’ clinical status, providing feedback to primary care providers, ensuring that guideline-based services are provided, and conveying interest and concern to depressed patients who may feel helpless and even hopeless.²⁰

Conclusions

Modified treatment algorithms can enhance the effectiveness of available interventions, but questions about their compatibility with the present organizational structure of primary care practice ultimately must be confronted and resolved. The manner in which this occurs will determine whether research findings from the laboratory can be transferred to the ambulatory medical setting in ways that benefit the large numbers of depressed patients in primary care settings.

In the following article, we summarize the presentations given at a conference on improving care for depression in organized health care systems. The 68 conference participants included mental health services researchers, persons responsible for improving care for depression in large health care systems, representatives of the National Institute of Mental Health, and foundation representatives. The specific aims of the conference were to: (1) consider what depression interventions are ready for dissemination in large organized health care systems; (2) identify unanswered questions concerning their effectiveness and cost-effectiveness; and (3) identify critical next steps to accelerate dissemination. (J Fam Pract 2001; 50:530-31)

Framework

The organizing framework for the conference was provided by the Model for Improving Chronic Illness Care.¹ This model identifies the need for health care systems change in 6 areas: (1) community resources and policies, including community programs and policies (eg, insurance benefits); (2) organization of care, including effective leadership for systems change and incentives for improved care; (3) self-management support consisting of a collaborative process between patients and providers to define problems, set priorities, establish goals, create treatment plans, and solve problems along the way²; (4) delivery system design, including clear delegation of roles and responsibilities from the physician to other professionals and systems for preventive services and active follow-up^3-5; (5) decision support, including guidelines, reminders, provider education, and appropriate input from relevant medical specialties^6,7; and (6) clinical information systems that provide timely information about patients with chronic conditions in support of guideline-based care.^8,9 Comprehensive descriptions of the model are available elsewhere in the literature,¹⁴ and at the Web site of the Robert Wood Johnson Foundation National Program for Improving Chronic Illness Care (www.improvingchroniccare.org). Conference participants considered whether general approaches to improving chronic illness care were relevant to depression.

Presentations

Nine major randomized controlled trials evaluating delivery of depression treatments were presented at the conference. Barrett^10,11 presented a trial evaluating antidepressant medications and Problem-Solving Therapy (PST) for patients with minor depression or dysthymia that showed benefits for medications but not for PST. Mynors-Wallis¹² presented a comparison of pharmacologic treatment, PST, and their combination for major depression that found equal benefits of medications and PST, but no advantage to combined therapy. Miranda¹³ presented an evaluation of cognitive-behavioral therapy for low-income minority individuals, with and without social case management, that showed no added benefit of social case management. Katon¹⁴ presented a trial of a collaborative program for patients with major depression at risk for chronic depression that showed benefits relative to usual care. Wells¹⁵ presented a multisite evaluation of a program of nurse follow-up to improve pharmacologic management or cognitive-behavioral therapy by mental health specialists that found improvement in depression outcomes and reduced unemployment relative to usual care. Rost¹⁶ presented a trial of nurse case management for patients with major depression that found improved depression outcomes relative to usual care. Katzelnick¹⁷ presented a multicenter trial of step-wise case management of patients with high utilization and major depression that showed improved outcomes relative to usual care. Simon¹⁸ presented a trial comparing feedback on treatment adherence and outcomes to the primary care physician to a telephone case management program that found improved outcomes relative to usual care for case management but not for feedback alone. Hunkeler¹⁹ presented a trial of nurse telephone follow-up for patients with major depression that found improved outcomes relative to usual care.

Acknowledgments

This conference was supported by grants from the Robert Wood Johnson Foundation, the John A. Hartford Foundation, and the National Institute of Mental Health. We gratefully acknowledge Christian Helfrich for making conference arrangements.

In the following article, we summarize the presentations given at a conference on improving care for depression in organized health care systems. The 68 conference participants included mental health services researchers, persons responsible for improving care for depression in large health care systems, representatives of the National Institute of Mental Health, and foundation representatives. The specific aims of the conference were to: (1) consider what depression interventions are ready for dissemination in large organized health care systems; (2) identify unanswered questions concerning their effectiveness and cost-effectiveness; and (3) identify critical next steps to accelerate dissemination. (J Fam Pract 2001; 50:530-31)

Framework

The organizing framework for the conference was provided by the Model for Improving Chronic Illness Care.¹ This model identifies the need for health care systems change in 6 areas: (1) community resources and policies, including community programs and policies (eg, insurance benefits); (2) organization of care, including effective leadership for systems change and incentives for improved care; (3) self-management support consisting of a collaborative process between patients and providers to define problems, set priorities, establish goals, create treatment plans, and solve problems along the way²; (4) delivery system design, including clear delegation of roles and responsibilities from the physician to other professionals and systems for preventive services and active follow-up^3-5; (5) decision support, including guidelines, reminders, provider education, and appropriate input from relevant medical specialties^6,7; and (6) clinical information systems that provide timely information about patients with chronic conditions in support of guideline-based care.^8,9 Comprehensive descriptions of the model are available elsewhere in the literature,¹⁴ and at the Web site of the Robert Wood Johnson Foundation National Program for Improving Chronic Illness Care (www.improvingchroniccare.org). Conference participants considered whether general approaches to improving chronic illness care were relevant to depression.

Presentations

Nine major randomized controlled trials evaluating delivery of depression treatments were presented at the conference. Barrett^10,11 presented a trial evaluating antidepressant medications and Problem-Solving Therapy (PST) for patients with minor depression or dysthymia that showed benefits for medications but not for PST. Mynors-Wallis¹² presented a comparison of pharmacologic treatment, PST, and their combination for major depression that found equal benefits of medications and PST, but no advantage to combined therapy. Miranda¹³ presented an evaluation of cognitive-behavioral therapy for low-income minority individuals, with and without social case management, that showed no added benefit of social case management. Katon¹⁴ presented a trial of a collaborative program for patients with major depression at risk for chronic depression that showed benefits relative to usual care. Wells¹⁵ presented a multisite evaluation of a program of nurse follow-up to improve pharmacologic management or cognitive-behavioral therapy by mental health specialists that found improvement in depression outcomes and reduced unemployment relative to usual care. Rost¹⁶ presented a trial of nurse case management for patients with major depression that found improved depression outcomes relative to usual care. Katzelnick¹⁷ presented a multicenter trial of step-wise case management of patients with high utilization and major depression that showed improved outcomes relative to usual care. Simon¹⁸ presented a trial comparing feedback on treatment adherence and outcomes to the primary care physician to a telephone case management program that found improved outcomes relative to usual care for case management but not for feedback alone. Hunkeler¹⁹ presented a trial of nurse telephone follow-up for patients with major depression that found improved outcomes relative to usual care.

Acknowledgments

This conference was supported by grants from the Robert Wood Johnson Foundation, the John A. Hartford Foundation, and the National Institute of Mental Health. We gratefully acknowledge Christian Helfrich for making conference arrangements.

In the following article, we summarize the presentations given at a conference on improving care for depression in organized health care systems. The 68 conference participants included mental health services researchers, persons responsible for improving care for depression in large health care systems, representatives of the National Institute of Mental Health, and foundation representatives. The specific aims of the conference were to: (1) consider what depression interventions are ready for dissemination in large organized health care systems; (2) identify unanswered questions concerning their effectiveness and cost-effectiveness; and (3) identify critical next steps to accelerate dissemination. (J Fam Pract 2001; 50:530-31)

Framework

The organizing framework for the conference was provided by the Model for Improving Chronic Illness Care.¹ This model identifies the need for health care systems change in 6 areas: (1) community resources and policies, including community programs and policies (eg, insurance benefits); (2) organization of care, including effective leadership for systems change and incentives for improved care; (3) self-management support consisting of a collaborative process between patients and providers to define problems, set priorities, establish goals, create treatment plans, and solve problems along the way²; (4) delivery system design, including clear delegation of roles and responsibilities from the physician to other professionals and systems for preventive services and active follow-up^3-5; (5) decision support, including guidelines, reminders, provider education, and appropriate input from relevant medical specialties^6,7; and (6) clinical information systems that provide timely information about patients with chronic conditions in support of guideline-based care.^8,9 Comprehensive descriptions of the model are available elsewhere in the literature,¹⁴ and at the Web site of the Robert Wood Johnson Foundation National Program for Improving Chronic Illness Care (www.improvingchroniccare.org). Conference participants considered whether general approaches to improving chronic illness care were relevant to depression.

Presentations

Nine major randomized controlled trials evaluating delivery of depression treatments were presented at the conference. Barrett^10,11 presented a trial evaluating antidepressant medications and Problem-Solving Therapy (PST) for patients with minor depression or dysthymia that showed benefits for medications but not for PST. Mynors-Wallis¹² presented a comparison of pharmacologic treatment, PST, and their combination for major depression that found equal benefits of medications and PST, but no advantage to combined therapy. Miranda¹³ presented an evaluation of cognitive-behavioral therapy for low-income minority individuals, with and without social case management, that showed no added benefit of social case management. Katon¹⁴ presented a trial of a collaborative program for patients with major depression at risk for chronic depression that showed benefits relative to usual care. Wells¹⁵ presented a multisite evaluation of a program of nurse follow-up to improve pharmacologic management or cognitive-behavioral therapy by mental health specialists that found improvement in depression outcomes and reduced unemployment relative to usual care. Rost¹⁶ presented a trial of nurse case management for patients with major depression that found improved depression outcomes relative to usual care. Katzelnick¹⁷ presented a multicenter trial of step-wise case management of patients with high utilization and major depression that showed improved outcomes relative to usual care. Simon¹⁸ presented a trial comparing feedback on treatment adherence and outcomes to the primary care physician to a telephone case management program that found improved outcomes relative to usual care for case management but not for feedback alone. Hunkeler¹⁹ presented a trial of nurse telephone follow-up for patients with major depression that found improved outcomes relative to usual care.

Acknowledgments

This conference was supported by grants from the Robert Wood Johnson Foundation, the John A. Hartford Foundation, and the National Institute of Mental Health. We gratefully acknowledge Christian Helfrich for making conference arrangements.

Irritable bowel syndrome (IBS) is a gastrointestinal problem estimated to affect 9% to 22% of the US population.^1-4 Recent studies suggest that although IBS symptoms are frequent in the community population,⁵ and those who do often discontinue care⁶ after receiving the diagnosis, despite continuing symptoms.⁷

As a diagnosis of exclusion with no defined physiologic basis, IBS is not well understood by physicians or patients and often not considered significant by the medical community.^8-10 There are no specific treatments for IBS and no evidencebased treatment guidelines.^11,12 This chronic condition has periodic exacerbation over many years that may adversely affect quality of life, especially in the areas of social and sexual functioning,¹³ events best studied from the patients’ perspective. However, few studies of IBS include this important perspective, and those that do usually only sample patients attending gastroenterology clinics, a very small minority of all IBS patients.

To better understand the patient’s view of the longitudinal nature of IBS, the factors that have an impact on symptom initiation and severity, and the needs of IBS suffers, we held community-based focus groups of people with diagnosed IBS. We attempted to determine the types of symptoms they experience, the chronicity and nature of episodic events, accommodations in lifestyles, successful treatments used, and IBS patients’ perceptions concerning their health and health care. This information should be useful to primary care physicians who provide the majority of care for people with IBS.

Methods

We performed a qualitative study using focus groups of volunteers previously diagnosed with IBS. Focus groups were held between September 15 and December 1, 1999. The subjects were identified by posted notices in primary care areas of a multispecialty private practice and through announcements in the local newspaper and at the local IBS support group. We used a structured interview guide reflecting the research questions to elicit responses from participants. During focus group discussions, the moderator used broad open-ended questions that began with low-intensity issues and moved toward higher-intensity issues. Specific probes were provided and modified after each session to incorporate findings from previous group discussions in an iterative process. The questions were initially tested using subjects from the local IBS support group to assure the questions’ comprehensibility and comprehensiveness.

The focus group discussion concentrated on 5 areas of interest: (1) description of IBS symptoms and episodes, (2) episode triggers and factors affecting the duration of episodes, (3) treatments used, (4) lifestyle changes due to IBS, and (5) experiences interfacing with the health care system. These areas of interest were selected on the basis of the existing literature^14-16 and gaps in the literature identified from the researchers’ clinical experience and previous work on IBS.^6,7,15*

Two trained research nurse co-facilitators led each focus group discussion. The master’s-prepared psychiatric nurse with extensive experience in group work led the group discussion, while a registered nurse experienced in focus groups served as an observer and recorder of content and nonverbal behavior. At the focus group meetings, the participants were welcomed and introduced using only first names. Time was spent informing participants of the group’s purpose and length, confidentiality issues and audiotaping, and that there were no right or wrong answers. It was stressed that the purpose of the sessions was to learn about participants’ experiences and opinions and not to attempt to reach a consensus about IBS. The 7 groups of 5 to 10 volunteers each met for 1 hour. After participants consented, the conversation was audiotaped and later transcribed verbatim. At the end of the session, the participants received a $25 honorarium for their time and travel.

Data Analysis

The 2 facilitators met immediately after each focus group to discuss the content of the discussion and to modify probes as necessary. The facilitators and one of the physician coinvestigators (BPY) initially developed summaries of responses to the questions of interest. They then read and reread the typed transcripts and group notes, sifting through the information, highlighting key statements, and looking for patterns and connections to identify dominant themes or concerns. This is known as immersion/crystallization.¹⁷ The researchers then met to review written emerging themes. In this process, each interpreter made connections of ideas and themes through an ongoing spiral of first organizing the data alone and then by corroborating it with the other interpreters. The process of crystallization occurred after several months of deliberations. We present summaries of answers to questions of interest and participant quotations to illustrate dominant themes that emerged in the crystallization process.

Results

Fifty-one volunteers participated (43 women and 8 men; age range=16-72 years). The occupations of the participants included: student, cook, teacher, health aide, musician, self-employed, office worker, chief executive officer of a small company, retired person, and clergy. Most participants were primary care patients and described being followed up medically by their usual physician for IBS symptoms before and after diagnosis.

Symptoms. Abdominal cramping and diarrhea caused most people with IBS to seek medical care. The symptoms often triggered a fear of serious disease, especially cancer. The symptoms were perceived as severe and limiting to the daily routine. Descriptions of abdominal pain were vivid, and participants consistently affirmed that the pain was not ordinary. The diarrhea was described as urgent and explosive. The sense of urgency that accompanies bouts of diarrhea and the inability to plan the day without taking into consideration the need for proximity to a bathroom was often frustrating. Bloating and constipation also occurred but were considered less disruptive of their lives. Participants often searched to find an etiology for their illness, linking the initial occurrence of symptoms with a specific event, such as a hernia or gallbladder operation, food poisoning, or even a cat bite.

IBS symptoms began at various ages, though most subjects were in their mid-teens to mid-30s when they sought help for the first time. Some were able to relate symptoms going back to early childhood. Many noted that constipation was a frequent problem during their preschool and elementary years and visits to physicians for constipation relief during childhood were not uncommon. Several participants said that they had been colicky babies and associated this experience with their IBS.

Episodes. The participants viewed IBS as a chronic disease that affects their lives on a daily basis. Although the symptoms were usually episodic, occurring weekly to monthly and lasting hours to a few days, anticipation of the next episode affected the person between episodes.

Stress and foods were considered common but unpredictable triggers of symptom episodes. Although all participants could identify possible stressors contributing to their symptoms, they universally believed that they had little control over these stressful occasions. Foods that could be eaten without adverse affect at one time triggered severe symptoms at other times. This variable response to potential triggers was a source of great frustration to the participants.

Treatment. Self-regulated symptomatic treatment was common: using an antidiarrheal drug when an episode of diarrhea began or fiber and bran in response to constipation. Prophylactic use of antidiarrheal medications during work or before important social situations was common. Use of antidepressant drugs for other reasons was noted to improve IBS symptoms. Despite lack of specific therapy, few participants used alternative health care providers or naturopathic remedies. IBS support groups were noted to be therapeutic.

Changes in lifestyle. Most participants reported they made significant changes in their work and social lives because of IBS. Overall, IBS was felt to decrease work productivity and adversely affect the quality of work due to large amounts of time spent in the bathroom. Work days were missed as often as once a month, mainly because of cramping or diarrhea.

Participants universally reported strained social situations because of the illness. Social situations become especially difficult when eating or traveling was involved. Most subjects said they simply avoid these situations by not participating in the social event or by not eating. Many refused social invitations for fear of an IBS episode and not having access to a bathroom. This causes great strain within the family unit.

Interface with the health care system. Medical help is sought in the diagnostic stages of IBS, but afterward most participants dealt with their symptoms on their own. Most felt that their physicians were not helpful with the illness. A sense of frustration with the medical community’s lack of empathic response, a lack of any helpful remedy, and a feeling of being taken lightly was commonplace.

Crystallization

Three overriding themes emerged from the focus groups responses: frustration, isolation, and a search for a niche in the health/sick role continuum Table 1.

Frustration. IBS sufferers displayed a sense of great frustration with the lack of control over their illness and daily lives. The unpredictability of IBS exacerbations, the severity of symptoms and the lack of understanding and empathy for their illness by family, coworkers, and physicians lead to feelings of disappointment, anger and frustration. Participants said: “It is triggered by anything, everything, nothing. It makes no sense”; “There is no control and people don’t understand that…no one can”; “All of a sudden it hits you. I’m afraid sometimes to get up and move”; “I haven’t noticed a pattern, and it’s frustrating”; “I think that part of the stress factor is the frustration of trying to figure out what’s coming next.”

The lack of an adequate medical explanation for IBS was frustrating for most, even though there was relief at the understanding that they did not have a serious medical disease. For some, the search for a cause of their IBS continued, and they explained their illness using life events that make sense for them. Participants viewed the illness as severe and greatly affecting their lives, and they believed that physicians did not give credence to the fact that they were ill. This perceived discrepancy caused them to see their physicians as unhelpful in dealing with IBS. One patient said: “There are medical people who think that it’s all in our head, and if they can’t come up with a diagnosis, it’s a wastebasket diagnosis.”

Frustration becomes a downward spiral triggering new episodes of symptoms leading to greater frustration and isolation. As one participant put it: “For me, my home…is not a very comfortable place, because my dad doesn’t understand this, and some of my family members don’t understand this.”

Isolation. The sense of isolation that participants felt was pervasive throughout the groups. Most felt apart from others, alone in their suffering, perceiving their symptoms as embarrassing and being unable or unwilling to discuss them with friends, families, and coworkers. Through group discussion, they were often surprised to learn of others’ similar episodes with IBS and felt an affirmation of their IBS experiences that they had not before experienced. One patient said: “I thought I was the only one in the whole world.”

Social events were difficult because of the perceived possibility of triggering an IBS episode and were often avoided because of frequent bathroom use, embarrassing occurrences of gas, abdominal pain, diarrhea, and even incontinence of stool. Isolation is further increased by the avoidance of social situations. Participants said: “I can’t go out to eat because I’m not sure what’s going to happen. I can’t go to family functions”; “You dread when somebody asks you to go somewhere, because you are afraid that you’re going to have a problem, and it’s going to affect everybody that you’re with”; “My family is always blaming me that I plan to have us not be able to go on vacation.”

Isolation in the workplace can also be enhanced when an IBS sufferer takes frequent bathroom breaks or must miss work because of abdominal pain. Coworkers’ unsympathetic responses toward the problem and frequent trips to the toilet increase the IBS suffers’ solitude. A participant said: “They (coworkers) complain a lot about me: ‘You’re just trying to get out of work’…that’s how they started teasing me, thinking I was just trying to skip out of work.”

Search for a niche. Participants viewed themselves as suffering from a chronic and sometimes debilitating illness. However, the message they received from the medical community was that there was nothing wrong with them. Nevertheless, the impact on daily life was considerable, and patients with IBS appeared to struggle to determine their place on the illness/wellness continuum. One patient said: “It is a waste of time to seek (help from a physician), because they say its just all in my head…they don’t want to deal with it.”

Even when feeling well, participants were concerned about the timing of their next episode. Thus, a great deal of energy was spent to avoid being ill. Participants said: “I try to get things stimulated in the morning so I can get it over with. It takes about an hour for the whole process because of the residual cramps,” and “I try to avoid certain foods.”

Discussion

These focus group volunteers were very willing to discuss their IBS symptoms and the ways their lives were affected by this illness. It seemed as if many were relieved to finally be able to discuss their problems and symptoms and to hear how others with similar difficulties fared. Quite a few were encouraged that their individual experiences with IBS problems paled in light of others’ difficulties with IBS.

People with IBS have experiences similar to those with other chronic conditions of uncertain etiology and ambiguous diagnostic criteria, such as fibromyalgia, chronic pain, and chronic fatigue. A phenomenalogic study of 22 women with fibromyalgia described 2 higher-order concepts called vulnerability and maintaining forces.¹⁸ The vulnerability was described in terms of feelings of helplessness similar to our theme of frustration. The maintaining forces dealt with ambiguous interaction with health professionals that described lack of recognition of the physical nature of this condition and symptoms. Other forces referred to family support and isolation. Studying a group of people with chronic fatigue, Clements and colleagues¹⁹ also reported an overlap of physical symptoms and stress over which sufferers have no control. They reported that coping with their condition requires avoiding activities and resultant isolation. Studying patients with chronic nonmalignant pain, Kelly and coworkers²⁰ described feelings of loss and the need to be understood and accepted that are similar to themes from our focus groups. The subjects from all 3 studies also affirmed the value of support group membership.

Our patients with IBS made specific suggestions to improve their interactions with physicians. More education about the nature of this illness at the time of diagnosis would be helpful to the IBS sufferer. Educating physicians about how IBS affects the lives of those who are plagued by this syndrome is paramount. An empathic response is warranted from health care workers, especially since this illness is so socially isolating. Because participants were frustrated by the social isolation of this illness, referral to an IBS support group could help decrease this sense of isolation and help the participants learn new coping methods. Although these focus groups were not therapy sessions, it was clear that many participants valued the group interaction and gained information about how others coped with their IBS. Most participants believed that their episodes were triggered largely by stress. They had difficulty determining how their stress could be minimized and, in general, most showed little understanding about stress management. Physicians should help their patients with IBS develop rational strategies for managing stress.

Limitations

The members of these focus groups were volunteers and may not be representative of the whole population. They may represent the more severe cases of IBS, those least likely to have found relief or those most dissatisfied with their health care. However, they were willing to call a nurse and participate in a hospital or clinicbased discussion. In addition, all subjects were white, which may not be representative of the United States culture at large or represent ethnic differences in the manifestation of the disease or the subjects’ solutions to their illness.

Conclusions

People with IBS view their illness as a chronic condition that is episodic in nature and triggered mainly by stressful events. They feel little control over the initiation, intensity, and duration of their episodes of illness, leading them to a feeling of frustration. They perceive their medical care providers as unsympathetic and of limited help in dealing with their disease. The gap between the patient and the physician’s perception of the nature, severity, and consequences of IBS appears to lead to frustration, isolation, and missed opportunities to improve the patient’s quality of life.

Related resources

For patients:

• IBS Self Help Group www.ibsgroup.org National IBS support group site. Offers information and chat rooms.
• HeliosHealth—Link to irritable bowel syndrome www.helioshealth.com/digestive/ibs Brochures, news and discussions.

For physicians:

• Advances in Irritable Bowel Syndrome www.conference-cast.com/ibs/ A lecture series that includes CME options.
• American Gastroenterology Association—Irritable Bowel Syndrome www.gastro.org/public/ibs.html This site addresses the types of IBS seen by GI specialty groups and includes guidelines for diagnosis.
• National Institute of Diabetes & Digestive & Kidney Diseases: Irritable Bowel Syndrome in Children http://www.niddk.nih.gov/health/digest/summary/ibskids/index.htm

Irritable bowel syndrome (IBS) is a gastrointestinal problem estimated to affect 9% to 22% of the US population.^1-4 Recent studies suggest that although IBS symptoms are frequent in the community population,⁵ and those who do often discontinue care⁶ after receiving the diagnosis, despite continuing symptoms.⁷

As a diagnosis of exclusion with no defined physiologic basis, IBS is not well understood by physicians or patients and often not considered significant by the medical community.^8-10 There are no specific treatments for IBS and no evidencebased treatment guidelines.^11,12 This chronic condition has periodic exacerbation over many years that may adversely affect quality of life, especially in the areas of social and sexual functioning,¹³ events best studied from the patients’ perspective. However, few studies of IBS include this important perspective, and those that do usually only sample patients attending gastroenterology clinics, a very small minority of all IBS patients.

To better understand the patient’s view of the longitudinal nature of IBS, the factors that have an impact on symptom initiation and severity, and the needs of IBS suffers, we held community-based focus groups of people with diagnosed IBS. We attempted to determine the types of symptoms they experience, the chronicity and nature of episodic events, accommodations in lifestyles, successful treatments used, and IBS patients’ perceptions concerning their health and health care. This information should be useful to primary care physicians who provide the majority of care for people with IBS.

Methods

We performed a qualitative study using focus groups of volunteers previously diagnosed with IBS. Focus groups were held between September 15 and December 1, 1999. The subjects were identified by posted notices in primary care areas of a multispecialty private practice and through announcements in the local newspaper and at the local IBS support group. We used a structured interview guide reflecting the research questions to elicit responses from participants. During focus group discussions, the moderator used broad open-ended questions that began with low-intensity issues and moved toward higher-intensity issues. Specific probes were provided and modified after each session to incorporate findings from previous group discussions in an iterative process. The questions were initially tested using subjects from the local IBS support group to assure the questions’ comprehensibility and comprehensiveness.

The focus group discussion concentrated on 5 areas of interest: (1) description of IBS symptoms and episodes, (2) episode triggers and factors affecting the duration of episodes, (3) treatments used, (4) lifestyle changes due to IBS, and (5) experiences interfacing with the health care system. These areas of interest were selected on the basis of the existing literature^14-16 and gaps in the literature identified from the researchers’ clinical experience and previous work on IBS.^6,7,15*

Two trained research nurse co-facilitators led each focus group discussion. The master’s-prepared psychiatric nurse with extensive experience in group work led the group discussion, while a registered nurse experienced in focus groups served as an observer and recorder of content and nonverbal behavior. At the focus group meetings, the participants were welcomed and introduced using only first names. Time was spent informing participants of the group’s purpose and length, confidentiality issues and audiotaping, and that there were no right or wrong answers. It was stressed that the purpose of the sessions was to learn about participants’ experiences and opinions and not to attempt to reach a consensus about IBS. The 7 groups of 5 to 10 volunteers each met for 1 hour. After participants consented, the conversation was audiotaped and later transcribed verbatim. At the end of the session, the participants received a $25 honorarium for their time and travel.

Data Analysis

The 2 facilitators met immediately after each focus group to discuss the content of the discussion and to modify probes as necessary. The facilitators and one of the physician coinvestigators (BPY) initially developed summaries of responses to the questions of interest. They then read and reread the typed transcripts and group notes, sifting through the information, highlighting key statements, and looking for patterns and connections to identify dominant themes or concerns. This is known as immersion/crystallization.¹⁷ The researchers then met to review written emerging themes. In this process, each interpreter made connections of ideas and themes through an ongoing spiral of first organizing the data alone and then by corroborating it with the other interpreters. The process of crystallization occurred after several months of deliberations. We present summaries of answers to questions of interest and participant quotations to illustrate dominant themes that emerged in the crystallization process.

Results

Fifty-one volunteers participated (43 women and 8 men; age range=16-72 years). The occupations of the participants included: student, cook, teacher, health aide, musician, self-employed, office worker, chief executive officer of a small company, retired person, and clergy. Most participants were primary care patients and described being followed up medically by their usual physician for IBS symptoms before and after diagnosis.

Symptoms. Abdominal cramping and diarrhea caused most people with IBS to seek medical care. The symptoms often triggered a fear of serious disease, especially cancer. The symptoms were perceived as severe and limiting to the daily routine. Descriptions of abdominal pain were vivid, and participants consistently affirmed that the pain was not ordinary. The diarrhea was described as urgent and explosive. The sense of urgency that accompanies bouts of diarrhea and the inability to plan the day without taking into consideration the need for proximity to a bathroom was often frustrating. Bloating and constipation also occurred but were considered less disruptive of their lives. Participants often searched to find an etiology for their illness, linking the initial occurrence of symptoms with a specific event, such as a hernia or gallbladder operation, food poisoning, or even a cat bite.

IBS symptoms began at various ages, though most subjects were in their mid-teens to mid-30s when they sought help for the first time. Some were able to relate symptoms going back to early childhood. Many noted that constipation was a frequent problem during their preschool and elementary years and visits to physicians for constipation relief during childhood were not uncommon. Several participants said that they had been colicky babies and associated this experience with their IBS.

Episodes. The participants viewed IBS as a chronic disease that affects their lives on a daily basis. Although the symptoms were usually episodic, occurring weekly to monthly and lasting hours to a few days, anticipation of the next episode affected the person between episodes.

Stress and foods were considered common but unpredictable triggers of symptom episodes. Although all participants could identify possible stressors contributing to their symptoms, they universally believed that they had little control over these stressful occasions. Foods that could be eaten without adverse affect at one time triggered severe symptoms at other times. This variable response to potential triggers was a source of great frustration to the participants.

Treatment. Self-regulated symptomatic treatment was common: using an antidiarrheal drug when an episode of diarrhea began or fiber and bran in response to constipation. Prophylactic use of antidiarrheal medications during work or before important social situations was common. Use of antidepressant drugs for other reasons was noted to improve IBS symptoms. Despite lack of specific therapy, few participants used alternative health care providers or naturopathic remedies. IBS support groups were noted to be therapeutic.

Changes in lifestyle. Most participants reported they made significant changes in their work and social lives because of IBS. Overall, IBS was felt to decrease work productivity and adversely affect the quality of work due to large amounts of time spent in the bathroom. Work days were missed as often as once a month, mainly because of cramping or diarrhea.

Participants universally reported strained social situations because of the illness. Social situations become especially difficult when eating or traveling was involved. Most subjects said they simply avoid these situations by not participating in the social event or by not eating. Many refused social invitations for fear of an IBS episode and not having access to a bathroom. This causes great strain within the family unit.

Interface with the health care system. Medical help is sought in the diagnostic stages of IBS, but afterward most participants dealt with their symptoms on their own. Most felt that their physicians were not helpful with the illness. A sense of frustration with the medical community’s lack of empathic response, a lack of any helpful remedy, and a feeling of being taken lightly was commonplace.

Crystallization

Three overriding themes emerged from the focus groups responses: frustration, isolation, and a search for a niche in the health/sick role continuum Table 1.

Frustration. IBS sufferers displayed a sense of great frustration with the lack of control over their illness and daily lives. The unpredictability of IBS exacerbations, the severity of symptoms and the lack of understanding and empathy for their illness by family, coworkers, and physicians lead to feelings of disappointment, anger and frustration. Participants said: “It is triggered by anything, everything, nothing. It makes no sense”; “There is no control and people don’t understand that…no one can”; “All of a sudden it hits you. I’m afraid sometimes to get up and move”; “I haven’t noticed a pattern, and it’s frustrating”; “I think that part of the stress factor is the frustration of trying to figure out what’s coming next.”

The lack of an adequate medical explanation for IBS was frustrating for most, even though there was relief at the understanding that they did not have a serious medical disease. For some, the search for a cause of their IBS continued, and they explained their illness using life events that make sense for them. Participants viewed the illness as severe and greatly affecting their lives, and they believed that physicians did not give credence to the fact that they were ill. This perceived discrepancy caused them to see their physicians as unhelpful in dealing with IBS. One patient said: “There are medical people who think that it’s all in our head, and if they can’t come up with a diagnosis, it’s a wastebasket diagnosis.”

Frustration becomes a downward spiral triggering new episodes of symptoms leading to greater frustration and isolation. As one participant put it: “For me, my home…is not a very comfortable place, because my dad doesn’t understand this, and some of my family members don’t understand this.”

Isolation. The sense of isolation that participants felt was pervasive throughout the groups. Most felt apart from others, alone in their suffering, perceiving their symptoms as embarrassing and being unable or unwilling to discuss them with friends, families, and coworkers. Through group discussion, they were often surprised to learn of others’ similar episodes with IBS and felt an affirmation of their IBS experiences that they had not before experienced. One patient said: “I thought I was the only one in the whole world.”

Social events were difficult because of the perceived possibility of triggering an IBS episode and were often avoided because of frequent bathroom use, embarrassing occurrences of gas, abdominal pain, diarrhea, and even incontinence of stool. Isolation is further increased by the avoidance of social situations. Participants said: “I can’t go out to eat because I’m not sure what’s going to happen. I can’t go to family functions”; “You dread when somebody asks you to go somewhere, because you are afraid that you’re going to have a problem, and it’s going to affect everybody that you’re with”; “My family is always blaming me that I plan to have us not be able to go on vacation.”

Isolation in the workplace can also be enhanced when an IBS sufferer takes frequent bathroom breaks or must miss work because of abdominal pain. Coworkers’ unsympathetic responses toward the problem and frequent trips to the toilet increase the IBS suffers’ solitude. A participant said: “They (coworkers) complain a lot about me: ‘You’re just trying to get out of work’…that’s how they started teasing me, thinking I was just trying to skip out of work.”

Search for a niche. Participants viewed themselves as suffering from a chronic and sometimes debilitating illness. However, the message they received from the medical community was that there was nothing wrong with them. Nevertheless, the impact on daily life was considerable, and patients with IBS appeared to struggle to determine their place on the illness/wellness continuum. One patient said: “It is a waste of time to seek (help from a physician), because they say its just all in my head…they don’t want to deal with it.”

Even when feeling well, participants were concerned about the timing of their next episode. Thus, a great deal of energy was spent to avoid being ill. Participants said: “I try to get things stimulated in the morning so I can get it over with. It takes about an hour for the whole process because of the residual cramps,” and “I try to avoid certain foods.”

Discussion

These focus group volunteers were very willing to discuss their IBS symptoms and the ways their lives were affected by this illness. It seemed as if many were relieved to finally be able to discuss their problems and symptoms and to hear how others with similar difficulties fared. Quite a few were encouraged that their individual experiences with IBS problems paled in light of others’ difficulties with IBS.

People with IBS have experiences similar to those with other chronic conditions of uncertain etiology and ambiguous diagnostic criteria, such as fibromyalgia, chronic pain, and chronic fatigue. A phenomenalogic study of 22 women with fibromyalgia described 2 higher-order concepts called vulnerability and maintaining forces.¹⁸ The vulnerability was described in terms of feelings of helplessness similar to our theme of frustration. The maintaining forces dealt with ambiguous interaction with health professionals that described lack of recognition of the physical nature of this condition and symptoms. Other forces referred to family support and isolation. Studying a group of people with chronic fatigue, Clements and colleagues¹⁹ also reported an overlap of physical symptoms and stress over which sufferers have no control. They reported that coping with their condition requires avoiding activities and resultant isolation. Studying patients with chronic nonmalignant pain, Kelly and coworkers²⁰ described feelings of loss and the need to be understood and accepted that are similar to themes from our focus groups. The subjects from all 3 studies also affirmed the value of support group membership.

Our patients with IBS made specific suggestions to improve their interactions with physicians. More education about the nature of this illness at the time of diagnosis would be helpful to the IBS sufferer. Educating physicians about how IBS affects the lives of those who are plagued by this syndrome is paramount. An empathic response is warranted from health care workers, especially since this illness is so socially isolating. Because participants were frustrated by the social isolation of this illness, referral to an IBS support group could help decrease this sense of isolation and help the participants learn new coping methods. Although these focus groups were not therapy sessions, it was clear that many participants valued the group interaction and gained information about how others coped with their IBS. Most participants believed that their episodes were triggered largely by stress. They had difficulty determining how their stress could be minimized and, in general, most showed little understanding about stress management. Physicians should help their patients with IBS develop rational strategies for managing stress.

Limitations

The members of these focus groups were volunteers and may not be representative of the whole population. They may represent the more severe cases of IBS, those least likely to have found relief or those most dissatisfied with their health care. However, they were willing to call a nurse and participate in a hospital or clinicbased discussion. In addition, all subjects were white, which may not be representative of the United States culture at large or represent ethnic differences in the manifestation of the disease or the subjects’ solutions to their illness.

Conclusions

People with IBS view their illness as a chronic condition that is episodic in nature and triggered mainly by stressful events. They feel little control over the initiation, intensity, and duration of their episodes of illness, leading them to a feeling of frustration. They perceive their medical care providers as unsympathetic and of limited help in dealing with their disease. The gap between the patient and the physician’s perception of the nature, severity, and consequences of IBS appears to lead to frustration, isolation, and missed opportunities to improve the patient’s quality of life.

Related resources

For patients:

• IBS Self Help Group www.ibsgroup.org National IBS support group site. Offers information and chat rooms.
• HeliosHealth—Link to irritable bowel syndrome www.helioshealth.com/digestive/ibs Brochures, news and discussions.

For physicians:

• Advances in Irritable Bowel Syndrome www.conference-cast.com/ibs/ A lecture series that includes CME options.
• American Gastroenterology Association—Irritable Bowel Syndrome www.gastro.org/public/ibs.html This site addresses the types of IBS seen by GI specialty groups and includes guidelines for diagnosis.
• National Institute of Diabetes & Digestive & Kidney Diseases: Irritable Bowel Syndrome in Children http://www.niddk.nih.gov/health/digest/summary/ibskids/index.htm

Irritable bowel syndrome (IBS) is a gastrointestinal problem estimated to affect 9% to 22% of the US population.^1-4 Recent studies suggest that although IBS symptoms are frequent in the community population,⁵ and those who do often discontinue care⁶ after receiving the diagnosis, despite continuing symptoms.⁷

As a diagnosis of exclusion with no defined physiologic basis, IBS is not well understood by physicians or patients and often not considered significant by the medical community.^8-10 There are no specific treatments for IBS and no evidencebased treatment guidelines.^11,12 This chronic condition has periodic exacerbation over many years that may adversely affect quality of life, especially in the areas of social and sexual functioning,¹³ events best studied from the patients’ perspective. However, few studies of IBS include this important perspective, and those that do usually only sample patients attending gastroenterology clinics, a very small minority of all IBS patients.

To better understand the patient’s view of the longitudinal nature of IBS, the factors that have an impact on symptom initiation and severity, and the needs of IBS suffers, we held community-based focus groups of people with diagnosed IBS. We attempted to determine the types of symptoms they experience, the chronicity and nature of episodic events, accommodations in lifestyles, successful treatments used, and IBS patients’ perceptions concerning their health and health care. This information should be useful to primary care physicians who provide the majority of care for people with IBS.

Methods

We performed a qualitative study using focus groups of volunteers previously diagnosed with IBS. Focus groups were held between September 15 and December 1, 1999. The subjects were identified by posted notices in primary care areas of a multispecialty private practice and through announcements in the local newspaper and at the local IBS support group. We used a structured interview guide reflecting the research questions to elicit responses from participants. During focus group discussions, the moderator used broad open-ended questions that began with low-intensity issues and moved toward higher-intensity issues. Specific probes were provided and modified after each session to incorporate findings from previous group discussions in an iterative process. The questions were initially tested using subjects from the local IBS support group to assure the questions’ comprehensibility and comprehensiveness.

The focus group discussion concentrated on 5 areas of interest: (1) description of IBS symptoms and episodes, (2) episode triggers and factors affecting the duration of episodes, (3) treatments used, (4) lifestyle changes due to IBS, and (5) experiences interfacing with the health care system. These areas of interest were selected on the basis of the existing literature^14-16 and gaps in the literature identified from the researchers’ clinical experience and previous work on IBS.^6,7,15*

Two trained research nurse co-facilitators led each focus group discussion. The master’s-prepared psychiatric nurse with extensive experience in group work led the group discussion, while a registered nurse experienced in focus groups served as an observer and recorder of content and nonverbal behavior. At the focus group meetings, the participants were welcomed and introduced using only first names. Time was spent informing participants of the group’s purpose and length, confidentiality issues and audiotaping, and that there were no right or wrong answers. It was stressed that the purpose of the sessions was to learn about participants’ experiences and opinions and not to attempt to reach a consensus about IBS. The 7 groups of 5 to 10 volunteers each met for 1 hour. After participants consented, the conversation was audiotaped and later transcribed verbatim. At the end of the session, the participants received a $25 honorarium for their time and travel.

Data Analysis

The 2 facilitators met immediately after each focus group to discuss the content of the discussion and to modify probes as necessary. The facilitators and one of the physician coinvestigators (BPY) initially developed summaries of responses to the questions of interest. They then read and reread the typed transcripts and group notes, sifting through the information, highlighting key statements, and looking for patterns and connections to identify dominant themes or concerns. This is known as immersion/crystallization.¹⁷ The researchers then met to review written emerging themes. In this process, each interpreter made connections of ideas and themes through an ongoing spiral of first organizing the data alone and then by corroborating it with the other interpreters. The process of crystallization occurred after several months of deliberations. We present summaries of answers to questions of interest and participant quotations to illustrate dominant themes that emerged in the crystallization process.

Results

Fifty-one volunteers participated (43 women and 8 men; age range=16-72 years). The occupations of the participants included: student, cook, teacher, health aide, musician, self-employed, office worker, chief executive officer of a small company, retired person, and clergy. Most participants were primary care patients and described being followed up medically by their usual physician for IBS symptoms before and after diagnosis.

Symptoms. Abdominal cramping and diarrhea caused most people with IBS to seek medical care. The symptoms often triggered a fear of serious disease, especially cancer. The symptoms were perceived as severe and limiting to the daily routine. Descriptions of abdominal pain were vivid, and participants consistently affirmed that the pain was not ordinary. The diarrhea was described as urgent and explosive. The sense of urgency that accompanies bouts of diarrhea and the inability to plan the day without taking into consideration the need for proximity to a bathroom was often frustrating. Bloating and constipation also occurred but were considered less disruptive of their lives. Participants often searched to find an etiology for their illness, linking the initial occurrence of symptoms with a specific event, such as a hernia or gallbladder operation, food poisoning, or even a cat bite.

IBS symptoms began at various ages, though most subjects were in their mid-teens to mid-30s when they sought help for the first time. Some were able to relate symptoms going back to early childhood. Many noted that constipation was a frequent problem during their preschool and elementary years and visits to physicians for constipation relief during childhood were not uncommon. Several participants said that they had been colicky babies and associated this experience with their IBS.

Episodes. The participants viewed IBS as a chronic disease that affects their lives on a daily basis. Although the symptoms were usually episodic, occurring weekly to monthly and lasting hours to a few days, anticipation of the next episode affected the person between episodes.

Stress and foods were considered common but unpredictable triggers of symptom episodes. Although all participants could identify possible stressors contributing to their symptoms, they universally believed that they had little control over these stressful occasions. Foods that could be eaten without adverse affect at one time triggered severe symptoms at other times. This variable response to potential triggers was a source of great frustration to the participants.

Treatment. Self-regulated symptomatic treatment was common: using an antidiarrheal drug when an episode of diarrhea began or fiber and bran in response to constipation. Prophylactic use of antidiarrheal medications during work or before important social situations was common. Use of antidepressant drugs for other reasons was noted to improve IBS symptoms. Despite lack of specific therapy, few participants used alternative health care providers or naturopathic remedies. IBS support groups were noted to be therapeutic.

Changes in lifestyle. Most participants reported they made significant changes in their work and social lives because of IBS. Overall, IBS was felt to decrease work productivity and adversely affect the quality of work due to large amounts of time spent in the bathroom. Work days were missed as often as once a month, mainly because of cramping or diarrhea.

Participants universally reported strained social situations because of the illness. Social situations become especially difficult when eating or traveling was involved. Most subjects said they simply avoid these situations by not participating in the social event or by not eating. Many refused social invitations for fear of an IBS episode and not having access to a bathroom. This causes great strain within the family unit.

Interface with the health care system. Medical help is sought in the diagnostic stages of IBS, but afterward most participants dealt with their symptoms on their own. Most felt that their physicians were not helpful with the illness. A sense of frustration with the medical community’s lack of empathic response, a lack of any helpful remedy, and a feeling of being taken lightly was commonplace.

Crystallization

Three overriding themes emerged from the focus groups responses: frustration, isolation, and a search for a niche in the health/sick role continuum Table 1.

Frustration. IBS sufferers displayed a sense of great frustration with the lack of control over their illness and daily lives. The unpredictability of IBS exacerbations, the severity of symptoms and the lack of understanding and empathy for their illness by family, coworkers, and physicians lead to feelings of disappointment, anger and frustration. Participants said: “It is triggered by anything, everything, nothing. It makes no sense”; “There is no control and people don’t understand that…no one can”; “All of a sudden it hits you. I’m afraid sometimes to get up and move”; “I haven’t noticed a pattern, and it’s frustrating”; “I think that part of the stress factor is the frustration of trying to figure out what’s coming next.”

The lack of an adequate medical explanation for IBS was frustrating for most, even though there was relief at the understanding that they did not have a serious medical disease. For some, the search for a cause of their IBS continued, and they explained their illness using life events that make sense for them. Participants viewed the illness as severe and greatly affecting their lives, and they believed that physicians did not give credence to the fact that they were ill. This perceived discrepancy caused them to see their physicians as unhelpful in dealing with IBS. One patient said: “There are medical people who think that it’s all in our head, and if they can’t come up with a diagnosis, it’s a wastebasket diagnosis.”

Frustration becomes a downward spiral triggering new episodes of symptoms leading to greater frustration and isolation. As one participant put it: “For me, my home…is not a very comfortable place, because my dad doesn’t understand this, and some of my family members don’t understand this.”

Isolation. The sense of isolation that participants felt was pervasive throughout the groups. Most felt apart from others, alone in their suffering, perceiving their symptoms as embarrassing and being unable or unwilling to discuss them with friends, families, and coworkers. Through group discussion, they were often surprised to learn of others’ similar episodes with IBS and felt an affirmation of their IBS experiences that they had not before experienced. One patient said: “I thought I was the only one in the whole world.”

Social events were difficult because of the perceived possibility of triggering an IBS episode and were often avoided because of frequent bathroom use, embarrassing occurrences of gas, abdominal pain, diarrhea, and even incontinence of stool. Isolation is further increased by the avoidance of social situations. Participants said: “I can’t go out to eat because I’m not sure what’s going to happen. I can’t go to family functions”; “You dread when somebody asks you to go somewhere, because you are afraid that you’re going to have a problem, and it’s going to affect everybody that you’re with”; “My family is always blaming me that I plan to have us not be able to go on vacation.”

Isolation in the workplace can also be enhanced when an IBS sufferer takes frequent bathroom breaks or must miss work because of abdominal pain. Coworkers’ unsympathetic responses toward the problem and frequent trips to the toilet increase the IBS suffers’ solitude. A participant said: “They (coworkers) complain a lot about me: ‘You’re just trying to get out of work’…that’s how they started teasing me, thinking I was just trying to skip out of work.”

Search for a niche. Participants viewed themselves as suffering from a chronic and sometimes debilitating illness. However, the message they received from the medical community was that there was nothing wrong with them. Nevertheless, the impact on daily life was considerable, and patients with IBS appeared to struggle to determine their place on the illness/wellness continuum. One patient said: “It is a waste of time to seek (help from a physician), because they say its just all in my head…they don’t want to deal with it.”

Even when feeling well, participants were concerned about the timing of their next episode. Thus, a great deal of energy was spent to avoid being ill. Participants said: “I try to get things stimulated in the morning so I can get it over with. It takes about an hour for the whole process because of the residual cramps,” and “I try to avoid certain foods.”

Discussion

These focus group volunteers were very willing to discuss their IBS symptoms and the ways their lives were affected by this illness. It seemed as if many were relieved to finally be able to discuss their problems and symptoms and to hear how others with similar difficulties fared. Quite a few were encouraged that their individual experiences with IBS problems paled in light of others’ difficulties with IBS.

People with IBS have experiences similar to those with other chronic conditions of uncertain etiology and ambiguous diagnostic criteria, such as fibromyalgia, chronic pain, and chronic fatigue. A phenomenalogic study of 22 women with fibromyalgia described 2 higher-order concepts called vulnerability and maintaining forces.¹⁸ The vulnerability was described in terms of feelings of helplessness similar to our theme of frustration. The maintaining forces dealt with ambiguous interaction with health professionals that described lack of recognition of the physical nature of this condition and symptoms. Other forces referred to family support and isolation. Studying a group of people with chronic fatigue, Clements and colleagues¹⁹ also reported an overlap of physical symptoms and stress over which sufferers have no control. They reported that coping with their condition requires avoiding activities and resultant isolation. Studying patients with chronic nonmalignant pain, Kelly and coworkers²⁰ described feelings of loss and the need to be understood and accepted that are similar to themes from our focus groups. The subjects from all 3 studies also affirmed the value of support group membership.

Our patients with IBS made specific suggestions to improve their interactions with physicians. More education about the nature of this illness at the time of diagnosis would be helpful to the IBS sufferer. Educating physicians about how IBS affects the lives of those who are plagued by this syndrome is paramount. An empathic response is warranted from health care workers, especially since this illness is so socially isolating. Because participants were frustrated by the social isolation of this illness, referral to an IBS support group could help decrease this sense of isolation and help the participants learn new coping methods. Although these focus groups were not therapy sessions, it was clear that many participants valued the group interaction and gained information about how others coped with their IBS. Most participants believed that their episodes were triggered largely by stress. They had difficulty determining how their stress could be minimized and, in general, most showed little understanding about stress management. Physicians should help their patients with IBS develop rational strategies for managing stress.

Limitations

The members of these focus groups were volunteers and may not be representative of the whole population. They may represent the more severe cases of IBS, those least likely to have found relief or those most dissatisfied with their health care. However, they were willing to call a nurse and participate in a hospital or clinicbased discussion. In addition, all subjects were white, which may not be representative of the United States culture at large or represent ethnic differences in the manifestation of the disease or the subjects’ solutions to their illness.

Conclusions

People with IBS view their illness as a chronic condition that is episodic in nature and triggered mainly by stressful events. They feel little control over the initiation, intensity, and duration of their episodes of illness, leading them to a feeling of frustration. They perceive their medical care providers as unsympathetic and of limited help in dealing with their disease. The gap between the patient and the physician’s perception of the nature, severity, and consequences of IBS appears to lead to frustration, isolation, and missed opportunities to improve the patient’s quality of life.

Related resources

For patients:

• IBS Self Help Group www.ibsgroup.org National IBS support group site. Offers information and chat rooms.
• HeliosHealth—Link to irritable bowel syndrome www.helioshealth.com/digestive/ibs Brochures, news and discussions.

For physicians:

• Advances in Irritable Bowel Syndrome www.conference-cast.com/ibs/ A lecture series that includes CME options.
• American Gastroenterology Association—Irritable Bowel Syndrome www.gastro.org/public/ibs.html This site addresses the types of IBS seen by GI specialty groups and includes guidelines for diagnosis.
• National Institute of Diabetes & Digestive & Kidney Diseases: Irritable Bowel Syndrome in Children http://www.niddk.nih.gov/health/digest/summary/ibskids/index.htm

Obesity is a common condition with associated risks of morbidity and mortality.^1,2 Recent studies suggest that rates of obesity and overweight are increasing in the United States.^3,4 Primary care physicians are struggling to incorporate weight management issues into their daily practices and often do not find the time to discuss this issue with their patients.^5-14 In recent years, obesity experts have developed guidelines to help physicians assist their obese patients with losing weight.^15-18 As primary care physicians look for ways to implement these guidelines with positive results, an understanding of patients’ weight management experiences and expectations in the physician-patient setting will be helpful. Unfortunately, few studies have addressed this topic.^19-21 With this study, we hoped to gain greater understanding of weight management issues from the patient’s point of view.

Methods

Research Environment

Two primary care practices in San Francisco were used as the study settings; both are affiliated with the University of California, San Francisco (UCSF): (1) the UCSF Medical Group at Parnassus Heights (which included 5 family physicians and 2 general internists), and (2) the UCSF Medical Group at Lakeshore (which included 4 family physicians and 1 general internist). Approximately two thirds of the patients in these practices had managed care insurance; 10% had Medicare; and 10% had Medicaid. The remaining patients paid for health care with other forms of fee-for-service reimbursement.

Sample Selection

During June and July 1997, we approached 214 adult patients at Parnassus Heights and 196 patients at Lakeshore to ask them to participate in our study. These 410 participants were drawn from consecutive samples of patients who were in the waiting room for a primary care visit. Pregnant patients, those younger than 18 years, and those who could not read English were excluded from the survey. We enrolled 366 (89%) of the patients approached.

Research Instrument

The participants completed an anonymous survey containing 15 general questions that required either yes or no responses or a selection from a list of choices. Only patients reporting that they needed to lose weight were asked how their weight affects them, their weight loss goals, and the role of their primary care physicians in helping them achieve these goals. However, all patients were asked if they had discussed their weight with their physicians in the past and whether they would feel comfortable discussing it with them in the future. Also, all patients were asked to describe the interventions used by their physicians in the past and those they would like to receive in the future. Finally, all patients were asked to provide demographic information and comorbidities. They were then directed to remove their shoes and coats and had weight and height recorded on a calibrated scale.

Statistical Analyses

We used the patients’ measurements to calculate body mass index (BMI=weight in kg/height in m²). The patients were then divided into 3 groups: those who were not overweight (BMI <25), those who were overweight (BMI=25-30), and those who were obese (BMI >30). The chi-square statistic was used to determine statistically significant associations between these groups and their responses to survey questions. We used analysis of variance testing to compare means between groups. The McNemar test was used to determine whether there was a statistically significant difference between patients’ past experiences and future preferences for intervention by their primary care physician. Finally, we did stepwise logistic regression analyses to determine predictors of patients reporting any communication with their physicians about their weight, and to determine predictors of overweight and obese patients reporting previous help with weight loss from their physicians. Odds ratios and 95% confidence intervals were determined. We conducted all analyses using SAS software.²²

Results

Patient Characteristics

The demographic characteristics of the patients are presented in Table 1. We found that 160 patients (44%) were not overweight, 101 (28%) were overweight, and 106 (29%) were obese. The patients in each group were similar in the number of visits to and the length of time with their primary care physicians. Obese patients were more likely to be African American or Latino and were less likely to be Asian American than patients who were not obese (P <.001). The prevalence of obesity-related comorbidities increased with higher BMI (P <.001). Ninety-seven percent of obese patients, 84% of overweight patients, and 39% of non-overweight patients felt they needed to lose weight (P <.001).

Prevalence of Physician-Patient Communication About Weight

We asked all patients if they had discussed their weight with their physicians in the past Table 2. Patients with higher BMIs were more likely to have had such a discussion. Those with lower BMIs were more likely to indicate that they had not needed help from their physicians with their weight in the past, and they were also more likely to indicate that they did not want such help in the future (P <.001 for all associations). Nearly all patients said they would be comfortable discussing weight with their physicians.

Attitudes of Patients Who Think They Need to Lose Weight

Table 3 shows the attitudes of patients who said they needed to lose weight. Most patients in all 3 categories indicated that weight loss was important to them and that their weight affected their happiness. However, patients with a higher BMI were more likely to state that their weight affects their health. Most patients in each group chose exercise and diet as the best methods of weight loss. However, patients with higher BMIs were more likely to want to be referred to a weight loss program. These patients were also more likely both to have felt that their physicians had helped in the past and that they could help in the future, and these patients had more ambitious weight loss goals (P <.001 for all associations).

What Was Done in the Past and What Patients Want in the Future

We asked all patients (regardless of whether they said they needed to lose weight) what their physicians had done about their weight in the past and what they would like from their physicians in the future Table 4. The most common experience reported was that physicians had not brought up weight. The patients reported a variety of other interventions that occurred, but none with a frequency greater than 20%. However, obese patients reported that their physicians intervened more. Although 33% of obese patients reported that their physicians had not brought up their weight, 48% of those patients said that they had been told to lose weight. The most common additional interventions reported by obese patients were: (1) a discussion of the health risks of obesity (31%), (2) exercise recommendations (30%), and (3) dietary advice (27%).

The most commonly cited future preferences for physician assistance of all patients were: (1) dietary advice (28%), (2) help setting realistic weight goals (27%), and (3) exercise recommendations (26%). The least-desired specific interventions were for their physicians to not bring up weight (12%) and for the physicians to say they do not have a weight problems (3%). Patients in all groups wanted substantially more future involvement by their physicians than they had experienced in the past, especially in setting realistic weight goals. Patients with higher BMIs were more likely to report both a higher level of intervention in the past and a desire for a higher level of intervention in the future. For example, although 13% of the obese patients stated that their physicians had helped them set realistic weight goals in the past, 46% said they would like their physicians to help them set realistic weight goals in the future, a difference of 33% (P <.001).

Predictors of Physician-Patient Communication About Weight

We did stepwise logistic regression analyses to determine predictors of physician-patient communication about weight. The logistic regression models offered the variables of Table 1 and patient BMI as candidates for consideration of statistical significance. For patients with BMIs less than 25 (not overweight), we found no significant predictors of such communication. For patients with BMIs greater than 25 (overweight or obese), we found that the diagnoses of diabetes (odds ratio [OR] =3.2; 95% confidence interval [CI], 1.2-9.2), high cholesterol (OR=2.6; 95% CI, 1.1-6.0), and depression (OR=2.4; 95% CI, 1.0-5.7) were predictive of patients reporting such a discussion had occurred in the past. Also, higher BMIs were also predictive of physician-patient communication about weight (OR=1.2; 95% CI, 1.1-1.3).

Interventions Used by Physicians Who Helped Patients Lose Weight

Of the 206 patients in our study who were either obese or overweight (BMI >25), 37 (18%) said that their current physicians had helped them lose weight in the past. We did a stepwise logistic regression analysis to learn more about what physician behaviors were predictive of this outcome. The logistic regression models offered the variables in Table 1 and patient BMI as candidates for consideration of statistical significance, as well as the items listed in Table 4. Physicians given credit by obese or overweight patients for helping them to lose weight in the past were more likely to have referred patients to weight loss groups and programs (OR=9.9; 95% CI, 1.9-59.6), made exercise recommendations (OR=9.3; 95% CI, 2.7-35.7), and helped patients understand the risks of their weight to their health (OR=4.5; 95% CI, 1.3-15.4).

Discussion

Nearly all patients in this diverse primary care population who might benefit from weight loss believed that they should lose weight. Most indicated that diet and exercise were the best weight loss methods for them. Also, 35% of obese patients and 20% of overweight patients believed that referral to a weight loss program could help them lose weight. Obese and overweight patients were generally receptive to the involvement of their primary care physicians with their weight concerns.

Unfortunately, only a minority of obese and overweight patients said they had discussed weight with their physicians in the past. We found that physicians targeted their communication about weight to patients with higher BMIs and obesity-related comorbidities. Similar findings have been reported by others in recent years.^9,12,14 Although this selective approach may have merit, it may also ignore the larger population of obese and overweight patients who may be receptive to a physician’s help with these issues. Also, it raises the question of whether primary care physicians are neglecting their responsibility to address weight management with patients before comorbidities develop.

The most common weight loss approaches used by physicians for obese patients were: (1) telling their patients to lose weight (48%) and (2) not bringing up the subject of their weight (33%). By far the most common weight loss approach physicians used for overweight patients was not bringing up their weight (64%). Also, although some patients said they wanted to be told by their physicians to lose weight (39% of obese patients and 13% of overweight patients), many wanted a more comprehensive approach, including dietary advice, exercise recommendations, and help in setting realistic weight goals. These results suggest that even when primary care physicians address weight issues with obese and overweight patients the discussion is not as intensive as many patients would like.

The diversity of responses given by obese and overweight patients suggests that physicians will need to tailor their advice to the individual needs of patients. However, we found that the small number of obese and overweight patients who credited their physicians with helping them lose weight were more likely to have received a few specific interventions. In particular, these patients were significantly more likely to have been referred to weight loss groups or programs, to have received exercise recommendations, and to have been given insights into the risks of their weight to their health. These might be important features for primary care physicians to consider incorporating into their approach to weight loss with their obese and overweight patients.

Perhaps the most surprising finding is that 39% of patients with BMIs less than 25 thought they should lose weight. Although a relatively small number of these patients may benefit from weight loss because of an unusually high waist-to-hip ratio or certain comorbidities that could confer increased risk,¹⁷ it is unlikely that more than a few of them would benefit medically from weight loss. Thus, physicians may need to work individually with these non-overweight patients to dispel myths and help them feel happier with their current weight.

Primary care physicians generally agree that prevention, identification, and treatment of weight problems and its comorbidities should be within their scope of practice.^23,24 Lack of time, training, teaching materials, staff support, and adequate reimbursement have been cited as common reasons they fail to address these issues more often in clinical practice.²⁵ Also, some physicians may not bring up weight issues for fear of negative patient reactions—fears that may have merit in some cases.²⁶ However, our study shows that the vast majority of patients are willing and even eager to discuss weight with their current primary care physicians.

Physicians may also neglect to bring up weight-related issues because they are uncertain whether such a discussion will have a positive impact on the health of their patients. A recent study confirmed that many obese and overweight patients who are seen in primary care are not ready to make the lifestyle changes needed for sustained weight loss.²¹ However, other studies suggest that physician encouragement can increase a patient’s readiness to make important lifestyle changes over time.^27,28 Recent studies also suggest that a physician’s advice to exercise can significantly improve activity levels over a several-month period.²⁹

Limitations

Our study was limited by its relatively small sample of patients in 2 group practices in San Francisco. However, the demographic findings were similar to those in larger studies, showing increasing rates of obesity in African Americans and Latinos and increasing rates of comorbidities such as diabetes mellitus, high blood pressure, and hyperlipidemia among those who were more overweight.² Thus, there is reason to believe that our study population has similarities to larger randomly selected primary care populations used to study some of these issues in the past. Another limitation is that all the data except for BMI were obtained from self-report of patients. However, although patients might underreport the level of communication they have with their physicians, patient self-report may ultimately be the most relevant measure of what communication is remembered by the patient.

Conclusions

The results of our study demonstrate that most patients, regardless of their weight, are open to greater physician involvement in weight management. This is important information for primary care physicians who want to address these needs. More research is needed to develop effective primary care approaches to weight management that are flexible and sensitive enough to meet the diverse needs of all patients.

Acknowledgments

This research was partially supported by a grant (#5D32PE19036-09) from the Health Resources Services Administration of the US Department of Health and Human Services to aid in the establishment of a department of family practice and by a grant from the California Academy of Family Physicians. We would also like to acknowledge Laurel Mellin, RD, who assisted with the development of the research instrument; Kim P. Truong, DO, who assisted with data collection; and Robert Wilson, PhD, who assisted with statistical analysis.

Related Resources

• National Heart, Lung and Blood Institute: clinical guidelines for obesity http://www.nhlbi.nih.gov/guidelines/obesity/ob_home.htm Includes the NIH guidelines on weight management and access to numerous clinical tools for physicians and patients.
• Shape Up America! http://www.shapeup.org/ A joint effort of Shape Up America! and the American Obesity Association, with numerous links that are relevant to clinicians and patients.

Obesity is a common condition with associated risks of morbidity and mortality.^1,2 Recent studies suggest that rates of obesity and overweight are increasing in the United States.^3,4 Primary care physicians are struggling to incorporate weight management issues into their daily practices and often do not find the time to discuss this issue with their patients.^5-14 In recent years, obesity experts have developed guidelines to help physicians assist their obese patients with losing weight.^15-18 As primary care physicians look for ways to implement these guidelines with positive results, an understanding of patients’ weight management experiences and expectations in the physician-patient setting will be helpful. Unfortunately, few studies have addressed this topic.^19-21 With this study, we hoped to gain greater understanding of weight management issues from the patient’s point of view.

Methods

Research Environment

Two primary care practices in San Francisco were used as the study settings; both are affiliated with the University of California, San Francisco (UCSF): (1) the UCSF Medical Group at Parnassus Heights (which included 5 family physicians and 2 general internists), and (2) the UCSF Medical Group at Lakeshore (which included 4 family physicians and 1 general internist). Approximately two thirds of the patients in these practices had managed care insurance; 10% had Medicare; and 10% had Medicaid. The remaining patients paid for health care with other forms of fee-for-service reimbursement.

Sample Selection

During June and July 1997, we approached 214 adult patients at Parnassus Heights and 196 patients at Lakeshore to ask them to participate in our study. These 410 participants were drawn from consecutive samples of patients who were in the waiting room for a primary care visit. Pregnant patients, those younger than 18 years, and those who could not read English were excluded from the survey. We enrolled 366 (89%) of the patients approached.

Research Instrument

The participants completed an anonymous survey containing 15 general questions that required either yes or no responses or a selection from a list of choices. Only patients reporting that they needed to lose weight were asked how their weight affects them, their weight loss goals, and the role of their primary care physicians in helping them achieve these goals. However, all patients were asked if they had discussed their weight with their physicians in the past and whether they would feel comfortable discussing it with them in the future. Also, all patients were asked to describe the interventions used by their physicians in the past and those they would like to receive in the future. Finally, all patients were asked to provide demographic information and comorbidities. They were then directed to remove their shoes and coats and had weight and height recorded on a calibrated scale.

Statistical Analyses

We used the patients’ measurements to calculate body mass index (BMI=weight in kg/height in m²). The patients were then divided into 3 groups: those who were not overweight (BMI <25), those who were overweight (BMI=25-30), and those who were obese (BMI >30). The chi-square statistic was used to determine statistically significant associations between these groups and their responses to survey questions. We used analysis of variance testing to compare means between groups. The McNemar test was used to determine whether there was a statistically significant difference between patients’ past experiences and future preferences for intervention by their primary care physician. Finally, we did stepwise logistic regression analyses to determine predictors of patients reporting any communication with their physicians about their weight, and to determine predictors of overweight and obese patients reporting previous help with weight loss from their physicians. Odds ratios and 95% confidence intervals were determined. We conducted all analyses using SAS software.²²

Results

Patient Characteristics

The demographic characteristics of the patients are presented in Table 1. We found that 160 patients (44%) were not overweight, 101 (28%) were overweight, and 106 (29%) were obese. The patients in each group were similar in the number of visits to and the length of time with their primary care physicians. Obese patients were more likely to be African American or Latino and were less likely to be Asian American than patients who were not obese (P <.001). The prevalence of obesity-related comorbidities increased with higher BMI (P <.001). Ninety-seven percent of obese patients, 84% of overweight patients, and 39% of non-overweight patients felt they needed to lose weight (P <.001).

Prevalence of Physician-Patient Communication About Weight

We asked all patients if they had discussed their weight with their physicians in the past Table 2. Patients with higher BMIs were more likely to have had such a discussion. Those with lower BMIs were more likely to indicate that they had not needed help from their physicians with their weight in the past, and they were also more likely to indicate that they did not want such help in the future (P <.001 for all associations). Nearly all patients said they would be comfortable discussing weight with their physicians.

Attitudes of Patients Who Think They Need to Lose Weight

Table 3 shows the attitudes of patients who said they needed to lose weight. Most patients in all 3 categories indicated that weight loss was important to them and that their weight affected their happiness. However, patients with a higher BMI were more likely to state that their weight affects their health. Most patients in each group chose exercise and diet as the best methods of weight loss. However, patients with higher BMIs were more likely to want to be referred to a weight loss program. These patients were also more likely both to have felt that their physicians had helped in the past and that they could help in the future, and these patients had more ambitious weight loss goals (P <.001 for all associations).

What Was Done in the Past and What Patients Want in the Future

We asked all patients (regardless of whether they said they needed to lose weight) what their physicians had done about their weight in the past and what they would like from their physicians in the future Table 4. The most common experience reported was that physicians had not brought up weight. The patients reported a variety of other interventions that occurred, but none with a frequency greater than 20%. However, obese patients reported that their physicians intervened more. Although 33% of obese patients reported that their physicians had not brought up their weight, 48% of those patients said that they had been told to lose weight. The most common additional interventions reported by obese patients were: (1) a discussion of the health risks of obesity (31%), (2) exercise recommendations (30%), and (3) dietary advice (27%).

The most commonly cited future preferences for physician assistance of all patients were: (1) dietary advice (28%), (2) help setting realistic weight goals (27%), and (3) exercise recommendations (26%). The least-desired specific interventions were for their physicians to not bring up weight (12%) and for the physicians to say they do not have a weight problems (3%). Patients in all groups wanted substantially more future involvement by their physicians than they had experienced in the past, especially in setting realistic weight goals. Patients with higher BMIs were more likely to report both a higher level of intervention in the past and a desire for a higher level of intervention in the future. For example, although 13% of the obese patients stated that their physicians had helped them set realistic weight goals in the past, 46% said they would like their physicians to help them set realistic weight goals in the future, a difference of 33% (P <.001).

Predictors of Physician-Patient Communication About Weight

We did stepwise logistic regression analyses to determine predictors of physician-patient communication about weight. The logistic regression models offered the variables of Table 1 and patient BMI as candidates for consideration of statistical significance. For patients with BMIs less than 25 (not overweight), we found no significant predictors of such communication. For patients with BMIs greater than 25 (overweight or obese), we found that the diagnoses of diabetes (odds ratio [OR] =3.2; 95% confidence interval [CI], 1.2-9.2), high cholesterol (OR=2.6; 95% CI, 1.1-6.0), and depression (OR=2.4; 95% CI, 1.0-5.7) were predictive of patients reporting such a discussion had occurred in the past. Also, higher BMIs were also predictive of physician-patient communication about weight (OR=1.2; 95% CI, 1.1-1.3).

Interventions Used by Physicians Who Helped Patients Lose Weight

Of the 206 patients in our study who were either obese or overweight (BMI >25), 37 (18%) said that their current physicians had helped them lose weight in the past. We did a stepwise logistic regression analysis to learn more about what physician behaviors were predictive of this outcome. The logistic regression models offered the variables in Table 1 and patient BMI as candidates for consideration of statistical significance, as well as the items listed in Table 4. Physicians given credit by obese or overweight patients for helping them to lose weight in the past were more likely to have referred patients to weight loss groups and programs (OR=9.9; 95% CI, 1.9-59.6), made exercise recommendations (OR=9.3; 95% CI, 2.7-35.7), and helped patients understand the risks of their weight to their health (OR=4.5; 95% CI, 1.3-15.4).

Discussion

Nearly all patients in this diverse primary care population who might benefit from weight loss believed that they should lose weight. Most indicated that diet and exercise were the best weight loss methods for them. Also, 35% of obese patients and 20% of overweight patients believed that referral to a weight loss program could help them lose weight. Obese and overweight patients were generally receptive to the involvement of their primary care physicians with their weight concerns.

Unfortunately, only a minority of obese and overweight patients said they had discussed weight with their physicians in the past. We found that physicians targeted their communication about weight to patients with higher BMIs and obesity-related comorbidities. Similar findings have been reported by others in recent years.^9,12,14 Although this selective approach may have merit, it may also ignore the larger population of obese and overweight patients who may be receptive to a physician’s help with these issues. Also, it raises the question of whether primary care physicians are neglecting their responsibility to address weight management with patients before comorbidities develop.

The most common weight loss approaches used by physicians for obese patients were: (1) telling their patients to lose weight (48%) and (2) not bringing up the subject of their weight (33%). By far the most common weight loss approach physicians used for overweight patients was not bringing up their weight (64%). Also, although some patients said they wanted to be told by their physicians to lose weight (39% of obese patients and 13% of overweight patients), many wanted a more comprehensive approach, including dietary advice, exercise recommendations, and help in setting realistic weight goals. These results suggest that even when primary care physicians address weight issues with obese and overweight patients the discussion is not as intensive as many patients would like.

The diversity of responses given by obese and overweight patients suggests that physicians will need to tailor their advice to the individual needs of patients. However, we found that the small number of obese and overweight patients who credited their physicians with helping them lose weight were more likely to have received a few specific interventions. In particular, these patients were significantly more likely to have been referred to weight loss groups or programs, to have received exercise recommendations, and to have been given insights into the risks of their weight to their health. These might be important features for primary care physicians to consider incorporating into their approach to weight loss with their obese and overweight patients.

Perhaps the most surprising finding is that 39% of patients with BMIs less than 25 thought they should lose weight. Although a relatively small number of these patients may benefit from weight loss because of an unusually high waist-to-hip ratio or certain comorbidities that could confer increased risk,¹⁷ it is unlikely that more than a few of them would benefit medically from weight loss. Thus, physicians may need to work individually with these non-overweight patients to dispel myths and help them feel happier with their current weight.

Primary care physicians generally agree that prevention, identification, and treatment of weight problems and its comorbidities should be within their scope of practice.^23,24 Lack of time, training, teaching materials, staff support, and adequate reimbursement have been cited as common reasons they fail to address these issues more often in clinical practice.²⁵ Also, some physicians may not bring up weight issues for fear of negative patient reactions—fears that may have merit in some cases.²⁶ However, our study shows that the vast majority of patients are willing and even eager to discuss weight with their current primary care physicians.

Physicians may also neglect to bring up weight-related issues because they are uncertain whether such a discussion will have a positive impact on the health of their patients. A recent study confirmed that many obese and overweight patients who are seen in primary care are not ready to make the lifestyle changes needed for sustained weight loss.²¹ However, other studies suggest that physician encouragement can increase a patient’s readiness to make important lifestyle changes over time.^27,28 Recent studies also suggest that a physician’s advice to exercise can significantly improve activity levels over a several-month period.²⁹

Limitations

Our study was limited by its relatively small sample of patients in 2 group practices in San Francisco. However, the demographic findings were similar to those in larger studies, showing increasing rates of obesity in African Americans and Latinos and increasing rates of comorbidities such as diabetes mellitus, high blood pressure, and hyperlipidemia among those who were more overweight.² Thus, there is reason to believe that our study population has similarities to larger randomly selected primary care populations used to study some of these issues in the past. Another limitation is that all the data except for BMI were obtained from self-report of patients. However, although patients might underreport the level of communication they have with their physicians, patient self-report may ultimately be the most relevant measure of what communication is remembered by the patient.

Conclusions

The results of our study demonstrate that most patients, regardless of their weight, are open to greater physician involvement in weight management. This is important information for primary care physicians who want to address these needs. More research is needed to develop effective primary care approaches to weight management that are flexible and sensitive enough to meet the diverse needs of all patients.

Acknowledgments

This research was partially supported by a grant (#5D32PE19036-09) from the Health Resources Services Administration of the US Department of Health and Human Services to aid in the establishment of a department of family practice and by a grant from the California Academy of Family Physicians. We would also like to acknowledge Laurel Mellin, RD, who assisted with the development of the research instrument; Kim P. Truong, DO, who assisted with data collection; and Robert Wilson, PhD, who assisted with statistical analysis.

Related Resources

• National Heart, Lung and Blood Institute: clinical guidelines for obesity http://www.nhlbi.nih.gov/guidelines/obesity/ob_home.htm Includes the NIH guidelines on weight management and access to numerous clinical tools for physicians and patients.
• Shape Up America! http://www.shapeup.org/ A joint effort of Shape Up America! and the American Obesity Association, with numerous links that are relevant to clinicians and patients.

Obesity is a common condition with associated risks of morbidity and mortality.^1,2 Recent studies suggest that rates of obesity and overweight are increasing in the United States.^3,4 Primary care physicians are struggling to incorporate weight management issues into their daily practices and often do not find the time to discuss this issue with their patients.^5-14 In recent years, obesity experts have developed guidelines to help physicians assist their obese patients with losing weight.^15-18 As primary care physicians look for ways to implement these guidelines with positive results, an understanding of patients’ weight management experiences and expectations in the physician-patient setting will be helpful. Unfortunately, few studies have addressed this topic.^19-21 With this study, we hoped to gain greater understanding of weight management issues from the patient’s point of view.

Methods

Research Environment

Two primary care practices in San Francisco were used as the study settings; both are affiliated with the University of California, San Francisco (UCSF): (1) the UCSF Medical Group at Parnassus Heights (which included 5 family physicians and 2 general internists), and (2) the UCSF Medical Group at Lakeshore (which included 4 family physicians and 1 general internist). Approximately two thirds of the patients in these practices had managed care insurance; 10% had Medicare; and 10% had Medicaid. The remaining patients paid for health care with other forms of fee-for-service reimbursement.

Sample Selection

During June and July 1997, we approached 214 adult patients at Parnassus Heights and 196 patients at Lakeshore to ask them to participate in our study. These 410 participants were drawn from consecutive samples of patients who were in the waiting room for a primary care visit. Pregnant patients, those younger than 18 years, and those who could not read English were excluded from the survey. We enrolled 366 (89%) of the patients approached.

Research Instrument

The participants completed an anonymous survey containing 15 general questions that required either yes or no responses or a selection from a list of choices. Only patients reporting that they needed to lose weight were asked how their weight affects them, their weight loss goals, and the role of their primary care physicians in helping them achieve these goals. However, all patients were asked if they had discussed their weight with their physicians in the past and whether they would feel comfortable discussing it with them in the future. Also, all patients were asked to describe the interventions used by their physicians in the past and those they would like to receive in the future. Finally, all patients were asked to provide demographic information and comorbidities. They were then directed to remove their shoes and coats and had weight and height recorded on a calibrated scale.

Statistical Analyses

We used the patients’ measurements to calculate body mass index (BMI=weight in kg/height in m²). The patients were then divided into 3 groups: those who were not overweight (BMI <25), those who were overweight (BMI=25-30), and those who were obese (BMI >30). The chi-square statistic was used to determine statistically significant associations between these groups and their responses to survey questions. We used analysis of variance testing to compare means between groups. The McNemar test was used to determine whether there was a statistically significant difference between patients’ past experiences and future preferences for intervention by their primary care physician. Finally, we did stepwise logistic regression analyses to determine predictors of patients reporting any communication with their physicians about their weight, and to determine predictors of overweight and obese patients reporting previous help with weight loss from their physicians. Odds ratios and 95% confidence intervals were determined. We conducted all analyses using SAS software.²²

Results

Patient Characteristics

The demographic characteristics of the patients are presented in Table 1. We found that 160 patients (44%) were not overweight, 101 (28%) were overweight, and 106 (29%) were obese. The patients in each group were similar in the number of visits to and the length of time with their primary care physicians. Obese patients were more likely to be African American or Latino and were less likely to be Asian American than patients who were not obese (P <.001). The prevalence of obesity-related comorbidities increased with higher BMI (P <.001). Ninety-seven percent of obese patients, 84% of overweight patients, and 39% of non-overweight patients felt they needed to lose weight (P <.001).

Prevalence of Physician-Patient Communication About Weight

We asked all patients if they had discussed their weight with their physicians in the past Table 2. Patients with higher BMIs were more likely to have had such a discussion. Those with lower BMIs were more likely to indicate that they had not needed help from their physicians with their weight in the past, and they were also more likely to indicate that they did not want such help in the future (P <.001 for all associations). Nearly all patients said they would be comfortable discussing weight with their physicians.

Attitudes of Patients Who Think They Need to Lose Weight

Table 3 shows the attitudes of patients who said they needed to lose weight. Most patients in all 3 categories indicated that weight loss was important to them and that their weight affected their happiness. However, patients with a higher BMI were more likely to state that their weight affects their health. Most patients in each group chose exercise and diet as the best methods of weight loss. However, patients with higher BMIs were more likely to want to be referred to a weight loss program. These patients were also more likely both to have felt that their physicians had helped in the past and that they could help in the future, and these patients had more ambitious weight loss goals (P <.001 for all associations).

What Was Done in the Past and What Patients Want in the Future

We asked all patients (regardless of whether they said they needed to lose weight) what their physicians had done about their weight in the past and what they would like from their physicians in the future Table 4. The most common experience reported was that physicians had not brought up weight. The patients reported a variety of other interventions that occurred, but none with a frequency greater than 20%. However, obese patients reported that their physicians intervened more. Although 33% of obese patients reported that their physicians had not brought up their weight, 48% of those patients said that they had been told to lose weight. The most common additional interventions reported by obese patients were: (1) a discussion of the health risks of obesity (31%), (2) exercise recommendations (30%), and (3) dietary advice (27%).

The most commonly cited future preferences for physician assistance of all patients were: (1) dietary advice (28%), (2) help setting realistic weight goals (27%), and (3) exercise recommendations (26%). The least-desired specific interventions were for their physicians to not bring up weight (12%) and for the physicians to say they do not have a weight problems (3%). Patients in all groups wanted substantially more future involvement by their physicians than they had experienced in the past, especially in setting realistic weight goals. Patients with higher BMIs were more likely to report both a higher level of intervention in the past and a desire for a higher level of intervention in the future. For example, although 13% of the obese patients stated that their physicians had helped them set realistic weight goals in the past, 46% said they would like their physicians to help them set realistic weight goals in the future, a difference of 33% (P <.001).

Predictors of Physician-Patient Communication About Weight

We did stepwise logistic regression analyses to determine predictors of physician-patient communication about weight. The logistic regression models offered the variables of Table 1 and patient BMI as candidates for consideration of statistical significance. For patients with BMIs less than 25 (not overweight), we found no significant predictors of such communication. For patients with BMIs greater than 25 (overweight or obese), we found that the diagnoses of diabetes (odds ratio [OR] =3.2; 95% confidence interval [CI], 1.2-9.2), high cholesterol (OR=2.6; 95% CI, 1.1-6.0), and depression (OR=2.4; 95% CI, 1.0-5.7) were predictive of patients reporting such a discussion had occurred in the past. Also, higher BMIs were also predictive of physician-patient communication about weight (OR=1.2; 95% CI, 1.1-1.3).

Interventions Used by Physicians Who Helped Patients Lose Weight

Of the 206 patients in our study who were either obese or overweight (BMI >25), 37 (18%) said that their current physicians had helped them lose weight in the past. We did a stepwise logistic regression analysis to learn more about what physician behaviors were predictive of this outcome. The logistic regression models offered the variables in Table 1 and patient BMI as candidates for consideration of statistical significance, as well as the items listed in Table 4. Physicians given credit by obese or overweight patients for helping them to lose weight in the past were more likely to have referred patients to weight loss groups and programs (OR=9.9; 95% CI, 1.9-59.6), made exercise recommendations (OR=9.3; 95% CI, 2.7-35.7), and helped patients understand the risks of their weight to their health (OR=4.5; 95% CI, 1.3-15.4).

Discussion

Nearly all patients in this diverse primary care population who might benefit from weight loss believed that they should lose weight. Most indicated that diet and exercise were the best weight loss methods for them. Also, 35% of obese patients and 20% of overweight patients believed that referral to a weight loss program could help them lose weight. Obese and overweight patients were generally receptive to the involvement of their primary care physicians with their weight concerns.

Unfortunately, only a minority of obese and overweight patients said they had discussed weight with their physicians in the past. We found that physicians targeted their communication about weight to patients with higher BMIs and obesity-related comorbidities. Similar findings have been reported by others in recent years.^9,12,14 Although this selective approach may have merit, it may also ignore the larger population of obese and overweight patients who may be receptive to a physician’s help with these issues. Also, it raises the question of whether primary care physicians are neglecting their responsibility to address weight management with patients before comorbidities develop.

The most common weight loss approaches used by physicians for obese patients were: (1) telling their patients to lose weight (48%) and (2) not bringing up the subject of their weight (33%). By far the most common weight loss approach physicians used for overweight patients was not bringing up their weight (64%). Also, although some patients said they wanted to be told by their physicians to lose weight (39% of obese patients and 13% of overweight patients), many wanted a more comprehensive approach, including dietary advice, exercise recommendations, and help in setting realistic weight goals. These results suggest that even when primary care physicians address weight issues with obese and overweight patients the discussion is not as intensive as many patients would like.

The diversity of responses given by obese and overweight patients suggests that physicians will need to tailor their advice to the individual needs of patients. However, we found that the small number of obese and overweight patients who credited their physicians with helping them lose weight were more likely to have received a few specific interventions. In particular, these patients were significantly more likely to have been referred to weight loss groups or programs, to have received exercise recommendations, and to have been given insights into the risks of their weight to their health. These might be important features for primary care physicians to consider incorporating into their approach to weight loss with their obese and overweight patients.

Perhaps the most surprising finding is that 39% of patients with BMIs less than 25 thought they should lose weight. Although a relatively small number of these patients may benefit from weight loss because of an unusually high waist-to-hip ratio or certain comorbidities that could confer increased risk,¹⁷ it is unlikely that more than a few of them would benefit medically from weight loss. Thus, physicians may need to work individually with these non-overweight patients to dispel myths and help them feel happier with their current weight.

Primary care physicians generally agree that prevention, identification, and treatment of weight problems and its comorbidities should be within their scope of practice.^23,24 Lack of time, training, teaching materials, staff support, and adequate reimbursement have been cited as common reasons they fail to address these issues more often in clinical practice.²⁵ Also, some physicians may not bring up weight issues for fear of negative patient reactions—fears that may have merit in some cases.²⁶ However, our study shows that the vast majority of patients are willing and even eager to discuss weight with their current primary care physicians.

Physicians may also neglect to bring up weight-related issues because they are uncertain whether such a discussion will have a positive impact on the health of their patients. A recent study confirmed that many obese and overweight patients who are seen in primary care are not ready to make the lifestyle changes needed for sustained weight loss.²¹ However, other studies suggest that physician encouragement can increase a patient’s readiness to make important lifestyle changes over time.^27,28 Recent studies also suggest that a physician’s advice to exercise can significantly improve activity levels over a several-month period.²⁹

Limitations

Our study was limited by its relatively small sample of patients in 2 group practices in San Francisco. However, the demographic findings were similar to those in larger studies, showing increasing rates of obesity in African Americans and Latinos and increasing rates of comorbidities such as diabetes mellitus, high blood pressure, and hyperlipidemia among those who were more overweight.² Thus, there is reason to believe that our study population has similarities to larger randomly selected primary care populations used to study some of these issues in the past. Another limitation is that all the data except for BMI were obtained from self-report of patients. However, although patients might underreport the level of communication they have with their physicians, patient self-report may ultimately be the most relevant measure of what communication is remembered by the patient.

Conclusions

The results of our study demonstrate that most patients, regardless of their weight, are open to greater physician involvement in weight management. This is important information for primary care physicians who want to address these needs. More research is needed to develop effective primary care approaches to weight management that are flexible and sensitive enough to meet the diverse needs of all patients.

Acknowledgments

This research was partially supported by a grant (#5D32PE19036-09) from the Health Resources Services Administration of the US Department of Health and Human Services to aid in the establishment of a department of family practice and by a grant from the California Academy of Family Physicians. We would also like to acknowledge Laurel Mellin, RD, who assisted with the development of the research instrument; Kim P. Truong, DO, who assisted with data collection; and Robert Wilson, PhD, who assisted with statistical analysis.

Related Resources

• National Heart, Lung and Blood Institute: clinical guidelines for obesity http://www.nhlbi.nih.gov/guidelines/obesity/ob_home.htm Includes the NIH guidelines on weight management and access to numerous clinical tools for physicians and patients.
• Shape Up America! http://www.shapeup.org/ A joint effort of Shape Up America! and the American Obesity Association, with numerous links that are relevant to clinicians and patients.

Obesity is a common problem, with significant health costs and associated morbidity; rates of overweight and obesity continue to increase in the United States^1,2 and Europe.³ In a typical general practice setting of 10,000 patients in the United Kingdom, approximately 1600 men and 1300 women could be expected to be overweight, and another 360 men and 700 women would need management of their obesity.⁴ Obesity predisposes patients to cardiovascular disease and hypertension,⁵ diabetes mellitus,⁶ hormone-related cancers,⁷ gallbladder disease,⁸ musculoskeletal disorders,⁹ sleep apnea,¹⁰ and generalized bodily pain.¹¹ It has a detrimental effect on social well-being,¹² decreases longevity,¹³ and imposes a tremendous financial and time burden on health care systems¹⁴; yet even a moderate amount of intentional weight loss (approximately 5%-10%) may significantly reduce the risk of mortality among subjects with medical complications of obesity.¹⁵

Increasingly, the major share of the primary effort to lower weight is falling on the general practice physician.¹⁶ Despite this, only 42% of obese patients seeking medical help are advised by their health professionals to lose weight; however, this telephone survey may suffer from a significant recall bias.¹⁷ Within the medical specialties, including family practice, internal medicine, and endocrinology, approximately 50% of physicians express interest in treating obesity in their practices.^16,18

Current treatment plans include mostly behavioral interventions, such as diet and exercise, along with behavior modification.¹⁹ These generally have a poor success rate, and in one study, long-term follow-up of weight maintenance after a 12-week very-low-calorie diet showed that patients regained the weight lost at a rate of 2.5% per month.²⁰ Although pharmacotherapy can be an important component of a weight loss program, safety issues have led to the withdrawal of some medications from the market. Studies with these have shown, however, that weight loss with pharmacotherapy can result in significant health benefits.^21-26

Sibutramine (Meridia; Knoll Pharmaceuticals, Nottingham, United Kingdom) is a new pharmacotherapy alternative. It is a serotonin-norepinephrine reuptake inhibitor (SNRI) that produces dose-related weight loss by enhancing postingestive satiety and by increasing energy expenditure.^21,27-29 Twelve months after weight reduction on a 4-week very-low-calorie diet, 75% of the patients taking sibutramine maintained 100% of their weight loss. This is a significantly greater percentage than the 42% who maintained 100% of weight loss in the nonsibutramine control group (P <.01).^30,31 Our study of sibutramine, carried out in a primary care general practice setting, was designed to determine the long-term efficacy, tolerability, and safety of dosages of 10 mg and 15 mg of sibutramine, which have been identified as effective and well tolerated in previous studies.^31,32 We also carried out analysis of improvements in risk factors resulting from 5% and 10% body weight loss.

Methods

The Nottinghamshire Independent Ethics Committee (Boots) approved our study, and all patients gave informed consent in writing. Our study was conducted in accordance with the Declaration of Helsinki (1983 revision).

Patients

We sought healthy male and female general practice patients, aged 18 to 65 years, who were mildly to moderately obese (body mass index [BMI] = 27 to 40 kg/m²), who had not lost more than 3 kg of weight during the previous 3 months, whose obesity was not of endocrine origin, and who did not have diabetes mellitus. All patients we enrolled were identified by primary care physicians. They had to have a seated pulse rate of 100 beats per minute or lower and a seated diastolic blood pressure of 100 mm Hg or lower. Hypertensive patients were included only if the condition had been stabilized with medication for 6 months. Patients receiving laxatives, anorectic agents, diuretics (except where stabilized for 6 months or more), bulking agents, antidepressants, or any other medication that in the opinion of the investigator might alter body weight were excluded. The patients completed the self-assessment Clinical Global Impressions questionnaire^32,33 and the Beck Depression Inventory. The objective of this was to identify patients who might be likely to have unusual weight changes in the context of the trial because of depression, and was not done for psychiatric evaluation. Those assessed by the investigator to be more than borderline depressed were excluded.

Trial Protocol

Because of the lack of a clear definition of obesity at the time of the initiation of our study and the lack of a clear definition of the patients for whom sibutramine treatment would be appropriate, protocol amendments permitted enrollment of a few patients with BMIs of 25 to 44 kg/m² (BMI Ž40 kg/m² is considered extremely obese). Depression and anxiety inventories were completed by a subset of patients at the study outset.³⁴ The investigator carried out a dietary inquiry for each eligible patient, and these patients were given standardized dietary advice and dietary advice sheets to be followed throughout the study. They were told to include certain foods in their diet each day: 12 oz of vegetables and fresh fruit; 6 oz of bread, cereal, potatoes, or rice; and 10 oz of skim milk; and they were told to substitute low-calorie foods such as fresh fruits and baked potatoes for sugary and fried foods such as chocolate and biscuits. The patients’ overall success in complying with dietary advice was assessed by the investigator using a 10-cm visual analog question scale labeled with opposite and extreme answers at either end (“easy to follow dietary advice” to “unable to follow dietary advice”).

Eligible patients then entered a 2-week single-blind placebo run-in period. At the end of the run-in period (month 0), those who still met the entry criteria entered the 12-month double-blind treatment phase of the study. Entry was restricted to those able to follow dietary advice as determined by the investigator. The patients were randomized in a double-blind fashion to once-daily treatment with placebo or sibutramine 10 mg or 15 mg dispensed in identical capsules, which were prepackaged and coded by the sponsor according to a computer-generated randomization list. The patients were assessed at monthly visits (month 0-month 12) during treatment and 1 week and 1 month after the completion of treatment. Patients withdrawing during treatment were assessed at that time and again 1 month after withdrawal. Beck Depression and State Anxiety inventories were performed on a subset of patients at the last treatment visit and at the 1-week follow-up visit to confirm lack of residual psychological changes due to medications.

At each visit, the investigator recorded the patient’s weight in indoor clothing. Waist and hip circumferences were measured at months 0, 6, and 12. Heart rate and blood pressure were recorded at every visit, and details of reported adverse events were recorded. Electrocardiograms and routine laboratory safety tests were performed at screening, 6 months, and 12 months.

Outcome Measures

The primary end point was the study outcome (measured on a categorical scale) on the basis of either the percentage weight loss at the end of the study or the reason for premature withdrawal.³⁵ The outcome measures were weight loss (kg), percentage weight loss, weight loss of at least 5% and at least 10%, change in BMI, and change in waist/hip ratio.

Statistical Analysis

Calculations were performed on the difference between treatment groups in weight change between baseline and month¹². Eighty-three patients completing treatment in each study group would be required to detect a difference in weight change of 4.0 kg with 90% power at the 5% significance level, assuming a standard deviation of 7.38 kg. Allowing for a forecast withdrawal rate of 45%, the desired sample size at enrollment was 184 patients per group or 552 for the entire study.

All statistical tests were two-tailed, with significance determined by reference to the 5% level. The null hypothesis was that the treatments were equivalent. The principal measure of efficacy was analyzed using the Kruskal-Wallis test,³⁶ including a factor for treatment group. Given that this result was significant at the 5% level, pairwise comparisons between the treatment groups were made by the large-sample normal approximation to the Wilcoxon rank sum test.³⁵

The methods we used to analyze differences between treatment groups were analysis of variance (ANOVA)³⁴ with factors for treatment group, center, and treatment group-by-center interaction for weight loss, waist and hip circumference, waist/hip ratio, vital signs, glucose and triglycerides, and ranked ANOVA for BMI, cholesterol, and uric acid. The differences between groups in the proportion of patients losing at least 5% and 10% of their baseline body weight were analyzed by logistic regression with factors for treatment group and center. All efficacy analyses were performed using the last observation carried forward (LOCF) to replace missing data. All data collected were included except for assessments performed more than 14 days after the last dose of study medication. After 14 days postdose, it was thought body weights could have changed significantly.

Results

Of 510 patients who entered the run-in period, 485 continued into the double-blind phase Figure 1. This occurred because of a lack of a gold-standard definition of obesity at the outset of this study; a protocol amendment allowed data from these patients to be included in the analysis. No clinically significant differences were noted between treatment groups at baseline with respect to demographic profile or characteristics of obesity Table 1. Four patients had BMIs at screening that fell out of the original trial protocol, all in the sibutramine 10-mg treatment group (1 patient <27 and 3 patients >40 kg/m²). Reanalysis of the data excluding these 4 patients was carried out, and no differences affecting the results, outcomes, or conclusions of our study were found.

Our analysis takes the patients who withdrew from the study into account.³² Pooled rates of withdrawal for any reason were similar in the 3 treatment groups: placebo, 83 patients (51%); sibutramine 10 mg, 79 (49%); and sibutramine 15 mg, 67 (42%). The completion rate for the 1-year study was 53%, which is generally consistent with weight loss programs.³⁷ Regarding the primary outcome analysis, fewer patients in the sibutramine groups than in the placebo group withdrew for lack of efficacy or adverse events. Using the log-rank test, there was no statistically significant difference between treatment groups in time to withdrawal, overall withdrawal rate, or withdrawal rate due to adverse events or lack of efficacy. Both sibutramine-treated groups had more patients in each category of weight loss above 5% (P=.001; Table 2. Pairwise comparisons showed a statistically significant difference versus placebo in favor of sibutramine 10 mg (P=.04) and sibutramine 15 mg (P <.001).

A significantly greater proportion of patients in the 2 sibutramine groups lost at least 5% and 10% of their baseline body weight Table 2. Mean weight reduction was significantly greater in the sibutramine groups than in the placebo group. Mean reductions in BMI reflected those in body weight and were significantly greater with 10 mg sibutramine (1.7 kg/m²) and 15 mg sibutramine (2.4 kg/m²) than with placebo (0.6 kg/m²; P <.001). The reduction in BMI was significantly greater with 15 mg sibutramine than 10-mg sibutramine (P <.05).

Patients treated with 10 mg or 15 mg sibutramine once daily lost more weight at each monthly assessment than those treated with placebo Figure 2. Weight loss was dose related, and maximal weight loss occurred by month 6 in all treatment groups. During the 4 weeks after treatment cessation (at whatever time this occurred), there were small weight increases in all treatment groups. Ten (6%) patients in the placebo group withdrew because of lack of efficacy compared with 5 (3%) in the sibutramine 10 mg group and 2 (1%) in the sibutramine 15-mg group.

Dose-related reductions in mean waist and hip circumferences achieved by month 6 were maintained at month 12. In the placebo group at month 12 (LOCF), the reductions were similar for waist and hip (2.4 cm and 2.6 cm, respectively), while in the 10- and 15-mg sibutramine groups, waist reduction was greater than hip reduction (6.4 cm and 3.8 cm at 10 mg and 7.4 cm and 5.2 cm at 15 mg). The reductions on active treatment were significantly greater than with placebo (P <.05 for overall comparison). The mean reduction in waist/hip ratio at month 12 (LOCF) was significantly greater with 10 mg sibutramine (-0.04) and 15 mg sibutramine (-0.03) than with placebo (-0.01; P <.05; least significant difference, .02).

There were statistically significant changes in triglycerides compared with placebo (+3.2%) at month 6 for the sibutramine 10- and 15-mg treatment groups (-10.8%, P <.05 and -10.0%, P <.01, respectively). Uric acid levels were also significantly reduced at month 6 for both sibutramine treatment groups (10 mg, -6.0% [P <.05] and 15 mg, -6.2% [P <.05] compared with placebo, -1.9%). Decrease in uric acid was the only laboratory-reported variable that was statistically significantly greater in the sibutramine-treated patients than in the placebo-treated patients at end point Table 3. These changes are more closely associated with disease end points than with patient outcomes.

Among adverse events occurring with a frequency of more than 5% in any treatment group, only dry mouth occurred significantly more frequently with 10 mg sibutramine (19 patients) or 15 mg sibutramine (21 patients) than with placebo (2 patients; P <.001). Eight serious adverse events led to early withdrawal during the double-blind phase of the study (2 in each of the sibutramine groups and 4 in the placebo group). Two were considered possibly related to sibutramine. The first was a 49-year-old woman with a history of epilepsy with no seizures during the previous 4 years, who was withdrawn after 126 days of treatment with 10 mg sibutramine because she had had 4 “drop attacks” within 2 weeks. The second was a 32-year-old woman who was withdrawn after 246 days of treatment with 15 mg sibutramine because of palpitations due to frequent ventricular ectopic beats. No evidence was found of any withdrawal effect during the follow-up period, when active treatment was over. Patient psychological status was evaluated both at the end of the double-blind treatment (63 patients completed all assessments) and at follow-up 1 week after stopping treatment (67 patients completed all assessments). There were no statistically or clinically significant differences between the treatment groups for change in either the Beck Depression Inventory or the State Anxiety Inventory studied from the end of the double-blind treatment period to the follow-up. All 3 groups showed a mean improvement in the depression inventory.

Mean changes in vital signs averaged over all post-baseline double-blind assessments adjusted for center and treatment-by-center are shown in Table 3. Differences in mean changes in blood pressure were not statistically significant for the sibutramine treatment groups from placebo, except for the mean increase in diastolic blood pressure in the sibutramine 10 mg treatment group Table 3. Statistically significant changes in pulse rate in the sibutramine 15 mg treatment group compared with the placebo treatment group were consistent with the mechanism of action of sibutramine as an SNRI.

Discussion

Weight loss, obesity, and overweight are important topics,¹⁶ and today an increasing proportion of obesity is first encountered in the family practice setting.³⁸ Our study shows that in the context of a family practice setting, a patient’s diet, along with sibutramine treatment, will cause more loss of weight than changes in diet alone. Increased levels of total serum cholesterol are associated with an increased risk of coronary heart disease, and although difficult to quantify, a reduction of 0.6 mmol per L has been observed to be associated with a 54% lower risk.³⁹ This was calculated in a retrospective analysis of results of several clinical studies. Patients in this study losing 10% of their body weight over 6 months taking sibutramine 10 mg once daily reduced serum cholesterol by 0.23 mmol per L and may contribute to an overall benefit for patients as part of a risk reduction program. Similarly, the recently recommended target for reduction of triglycerides is a reduction of 1.7 mmol per L.⁴⁰ In our study, a reduction of 0.63 mmol per L in triglycerides was accomplished among those patients who received sibutramine 10 mg who lost at least 10% of body weight. Mean changes in blood pressure observed during the 6 months of our study were not great enough to increase the category of patient risk for coronary heart disease.⁴¹ Although weight loss can be associated with decreased mortality,¹³ the extent to which a weight reduction of 10% can augment other medical approaches to risk reduction long-term is only now being defined.

In this 1-year placebo-controlled study carried out in the primary care setting, sibutramine given with dietary advice was significantly superior in all weight loss efficacy assessments to dietary advice alone (placebo). These efficacy assessments included the ranked weight reduction outcomes analysis. Sibutramine was consistently superior to placebo across all categories of weight loss in this ranked analysis that takes subject withdrawals into account. The results of this study confirm the findings previously reported.³⁰

The plateau in weight loss observed at 6 months and maintained while having treatment is consistent with observations in other long-term studies of drug treatment in uncomplicated obese patients.^26,37 On the basis of other studies of weight loss, the regain of weight after cessation of treatment is not unexpected. The rates of completion of the study, although low, probably reflect what might be expected to occur in a general practice setting and are consistent with completion rates of weight loss programs in general.³⁷

The response rate at the 5% and 10% weight-reduction thresholds, with changes in waist circumference and in waist/hip ratio, were chosen for analysis in part because weight loss at these levels may ameliorate obesity-related disease.^37-42 Sustained moderate weight loss may be expected to bring about a reduction in blood pressure, hypertension, and dyslipidemia.⁴³ The small mean changes in blood pressure seen in the sibutramine groups were not unexpected given the norepinephrine-reuptake inhibitory properties of sibutramine. In clinical practice, any clinically significant changes that may occur should be detected by routine monitoring of vital signs. Of note, even in the current setting, where approximately 8% of patients were being treated with concomitant therapies related to cardiovascular disease, no patients were withdrawn (0%) because of elevated blood pressure.

Limitations

There were limitations to the general applicability of our study results to family practice. The treatment was limited to those patients who first demonstrated they could adhere to a weight loss program. Although the family practitioner may have some background information concerning this capability for a candidate for sibutramine treatment, general documentation may not be available for all candidates. Also, only 53% of the randomized patients completed this 1-year weight loss study. In addition, generally patients who stop weight loss treatment gain weight; however, the Sibutramine Trial in Obesity Reduction and Maintenance⁴⁴ demonstrated that weight loss can be maintained safely with sibutramine for up to 2 years. Another limitation to applicability of the study is the detailed depression and anxiety profile carried out for some patients enrolled. Such analysis is not practical or needed before sibutramine use in a family practice setting, and this analysis is not included in recently initiated studies. This was a generally healthy obese patient population with few comorbidities, and they were not depressed or anxious. That may limit the ability to generalize these findings to obese patients overall.

Conclusions

Dosages of sibutramine 10 mg or 15 mg administered once daily in a primary care general practice setting to patients with mild to moderate uncomplicated obesity produced statistically and clinically significant weight reduction that was maintained over a 12-month treatment period. Sibutramine was safe and well tolerated at both 10-mg and 15-mg once-daily doses. Further studies in patients with medical complications of obesity are under way to evaluate the efficacy of sibutramine-induced weight loss on the comorbidities and risk factors associated with obesity.

Acknowledgments

The author would like to thank the Sibutramine Clinical Study 1047 Team for their contributions, especially J Hosie, I James, and SP Jones. Competing interests: This study was funded by Knoll Pharmaceuticals, the manufacturer of sibutramine.

Obesity is a common problem, with significant health costs and associated morbidity; rates of overweight and obesity continue to increase in the United States^1,2 and Europe.³ In a typical general practice setting of 10,000 patients in the United Kingdom, approximately 1600 men and 1300 women could be expected to be overweight, and another 360 men and 700 women would need management of their obesity.⁴ Obesity predisposes patients to cardiovascular disease and hypertension,⁵ diabetes mellitus,⁶ hormone-related cancers,⁷ gallbladder disease,⁸ musculoskeletal disorders,⁹ sleep apnea,¹⁰ and generalized bodily pain.¹¹ It has a detrimental effect on social well-being,¹² decreases longevity,¹³ and imposes a tremendous financial and time burden on health care systems¹⁴; yet even a moderate amount of intentional weight loss (approximately 5%-10%) may significantly reduce the risk of mortality among subjects with medical complications of obesity.¹⁵

Increasingly, the major share of the primary effort to lower weight is falling on the general practice physician.¹⁶ Despite this, only 42% of obese patients seeking medical help are advised by their health professionals to lose weight; however, this telephone survey may suffer from a significant recall bias.¹⁷ Within the medical specialties, including family practice, internal medicine, and endocrinology, approximately 50% of physicians express interest in treating obesity in their practices.^16,18

Current treatment plans include mostly behavioral interventions, such as diet and exercise, along with behavior modification.¹⁹ These generally have a poor success rate, and in one study, long-term follow-up of weight maintenance after a 12-week very-low-calorie diet showed that patients regained the weight lost at a rate of 2.5% per month.²⁰ Although pharmacotherapy can be an important component of a weight loss program, safety issues have led to the withdrawal of some medications from the market. Studies with these have shown, however, that weight loss with pharmacotherapy can result in significant health benefits.^21-26

Sibutramine (Meridia; Knoll Pharmaceuticals, Nottingham, United Kingdom) is a new pharmacotherapy alternative. It is a serotonin-norepinephrine reuptake inhibitor (SNRI) that produces dose-related weight loss by enhancing postingestive satiety and by increasing energy expenditure.^21,27-29 Twelve months after weight reduction on a 4-week very-low-calorie diet, 75% of the patients taking sibutramine maintained 100% of their weight loss. This is a significantly greater percentage than the 42% who maintained 100% of weight loss in the nonsibutramine control group (P <.01).^30,31 Our study of sibutramine, carried out in a primary care general practice setting, was designed to determine the long-term efficacy, tolerability, and safety of dosages of 10 mg and 15 mg of sibutramine, which have been identified as effective and well tolerated in previous studies.^31,32 We also carried out analysis of improvements in risk factors resulting from 5% and 10% body weight loss.

Methods

The Nottinghamshire Independent Ethics Committee (Boots) approved our study, and all patients gave informed consent in writing. Our study was conducted in accordance with the Declaration of Helsinki (1983 revision).

Patients

We sought healthy male and female general practice patients, aged 18 to 65 years, who were mildly to moderately obese (body mass index [BMI] = 27 to 40 kg/m²), who had not lost more than 3 kg of weight during the previous 3 months, whose obesity was not of endocrine origin, and who did not have diabetes mellitus. All patients we enrolled were identified by primary care physicians. They had to have a seated pulse rate of 100 beats per minute or lower and a seated diastolic blood pressure of 100 mm Hg or lower. Hypertensive patients were included only if the condition had been stabilized with medication for 6 months. Patients receiving laxatives, anorectic agents, diuretics (except where stabilized for 6 months or more), bulking agents, antidepressants, or any other medication that in the opinion of the investigator might alter body weight were excluded. The patients completed the self-assessment Clinical Global Impressions questionnaire^32,33 and the Beck Depression Inventory. The objective of this was to identify patients who might be likely to have unusual weight changes in the context of the trial because of depression, and was not done for psychiatric evaluation. Those assessed by the investigator to be more than borderline depressed were excluded.

Trial Protocol

Because of the lack of a clear definition of obesity at the time of the initiation of our study and the lack of a clear definition of the patients for whom sibutramine treatment would be appropriate, protocol amendments permitted enrollment of a few patients with BMIs of 25 to 44 kg/m² (BMI Ž40 kg/m² is considered extremely obese). Depression and anxiety inventories were completed by a subset of patients at the study outset.³⁴ The investigator carried out a dietary inquiry for each eligible patient, and these patients were given standardized dietary advice and dietary advice sheets to be followed throughout the study. They were told to include certain foods in their diet each day: 12 oz of vegetables and fresh fruit; 6 oz of bread, cereal, potatoes, or rice; and 10 oz of skim milk; and they were told to substitute low-calorie foods such as fresh fruits and baked potatoes for sugary and fried foods such as chocolate and biscuits. The patients’ overall success in complying with dietary advice was assessed by the investigator using a 10-cm visual analog question scale labeled with opposite and extreme answers at either end (“easy to follow dietary advice” to “unable to follow dietary advice”).

Eligible patients then entered a 2-week single-blind placebo run-in period. At the end of the run-in period (month 0), those who still met the entry criteria entered the 12-month double-blind treatment phase of the study. Entry was restricted to those able to follow dietary advice as determined by the investigator. The patients were randomized in a double-blind fashion to once-daily treatment with placebo or sibutramine 10 mg or 15 mg dispensed in identical capsules, which were prepackaged and coded by the sponsor according to a computer-generated randomization list. The patients were assessed at monthly visits (month 0-month 12) during treatment and 1 week and 1 month after the completion of treatment. Patients withdrawing during treatment were assessed at that time and again 1 month after withdrawal. Beck Depression and State Anxiety inventories were performed on a subset of patients at the last treatment visit and at the 1-week follow-up visit to confirm lack of residual psychological changes due to medications.

At each visit, the investigator recorded the patient’s weight in indoor clothing. Waist and hip circumferences were measured at months 0, 6, and 12. Heart rate and blood pressure were recorded at every visit, and details of reported adverse events were recorded. Electrocardiograms and routine laboratory safety tests were performed at screening, 6 months, and 12 months.

Outcome Measures

The primary end point was the study outcome (measured on a categorical scale) on the basis of either the percentage weight loss at the end of the study or the reason for premature withdrawal.³⁵ The outcome measures were weight loss (kg), percentage weight loss, weight loss of at least 5% and at least 10%, change in BMI, and change in waist/hip ratio.

Statistical Analysis

Calculations were performed on the difference between treatment groups in weight change between baseline and month¹². Eighty-three patients completing treatment in each study group would be required to detect a difference in weight change of 4.0 kg with 90% power at the 5% significance level, assuming a standard deviation of 7.38 kg. Allowing for a forecast withdrawal rate of 45%, the desired sample size at enrollment was 184 patients per group or 552 for the entire study.

All statistical tests were two-tailed, with significance determined by reference to the 5% level. The null hypothesis was that the treatments were equivalent. The principal measure of efficacy was analyzed using the Kruskal-Wallis test,³⁶ including a factor for treatment group. Given that this result was significant at the 5% level, pairwise comparisons between the treatment groups were made by the large-sample normal approximation to the Wilcoxon rank sum test.³⁵

The methods we used to analyze differences between treatment groups were analysis of variance (ANOVA)³⁴ with factors for treatment group, center, and treatment group-by-center interaction for weight loss, waist and hip circumference, waist/hip ratio, vital signs, glucose and triglycerides, and ranked ANOVA for BMI, cholesterol, and uric acid. The differences between groups in the proportion of patients losing at least 5% and 10% of their baseline body weight were analyzed by logistic regression with factors for treatment group and center. All efficacy analyses were performed using the last observation carried forward (LOCF) to replace missing data. All data collected were included except for assessments performed more than 14 days after the last dose of study medication. After 14 days postdose, it was thought body weights could have changed significantly.

Results

Of 510 patients who entered the run-in period, 485 continued into the double-blind phase Figure 1. This occurred because of a lack of a gold-standard definition of obesity at the outset of this study; a protocol amendment allowed data from these patients to be included in the analysis. No clinically significant differences were noted between treatment groups at baseline with respect to demographic profile or characteristics of obesity Table 1. Four patients had BMIs at screening that fell out of the original trial protocol, all in the sibutramine 10-mg treatment group (1 patient <27 and 3 patients >40 kg/m²). Reanalysis of the data excluding these 4 patients was carried out, and no differences affecting the results, outcomes, or conclusions of our study were found.

Our analysis takes the patients who withdrew from the study into account.³² Pooled rates of withdrawal for any reason were similar in the 3 treatment groups: placebo, 83 patients (51%); sibutramine 10 mg, 79 (49%); and sibutramine 15 mg, 67 (42%). The completion rate for the 1-year study was 53%, which is generally consistent with weight loss programs.³⁷ Regarding the primary outcome analysis, fewer patients in the sibutramine groups than in the placebo group withdrew for lack of efficacy or adverse events. Using the log-rank test, there was no statistically significant difference between treatment groups in time to withdrawal, overall withdrawal rate, or withdrawal rate due to adverse events or lack of efficacy. Both sibutramine-treated groups had more patients in each category of weight loss above 5% (P=.001; Table 2. Pairwise comparisons showed a statistically significant difference versus placebo in favor of sibutramine 10 mg (P=.04) and sibutramine 15 mg (P <.001).

A significantly greater proportion of patients in the 2 sibutramine groups lost at least 5% and 10% of their baseline body weight Table 2. Mean weight reduction was significantly greater in the sibutramine groups than in the placebo group. Mean reductions in BMI reflected those in body weight and were significantly greater with 10 mg sibutramine (1.7 kg/m²) and 15 mg sibutramine (2.4 kg/m²) than with placebo (0.6 kg/m²; P <.001). The reduction in BMI was significantly greater with 15 mg sibutramine than 10-mg sibutramine (P <.05).

Patients treated with 10 mg or 15 mg sibutramine once daily lost more weight at each monthly assessment than those treated with placebo Figure 2. Weight loss was dose related, and maximal weight loss occurred by month 6 in all treatment groups. During the 4 weeks after treatment cessation (at whatever time this occurred), there were small weight increases in all treatment groups. Ten (6%) patients in the placebo group withdrew because of lack of efficacy compared with 5 (3%) in the sibutramine 10 mg group and 2 (1%) in the sibutramine 15-mg group.

Dose-related reductions in mean waist and hip circumferences achieved by month 6 were maintained at month 12. In the placebo group at month 12 (LOCF), the reductions were similar for waist and hip (2.4 cm and 2.6 cm, respectively), while in the 10- and 15-mg sibutramine groups, waist reduction was greater than hip reduction (6.4 cm and 3.8 cm at 10 mg and 7.4 cm and 5.2 cm at 15 mg). The reductions on active treatment were significantly greater than with placebo (P <.05 for overall comparison). The mean reduction in waist/hip ratio at month 12 (LOCF) was significantly greater with 10 mg sibutramine (-0.04) and 15 mg sibutramine (-0.03) than with placebo (-0.01; P <.05; least significant difference, .02).

There were statistically significant changes in triglycerides compared with placebo (+3.2%) at month 6 for the sibutramine 10- and 15-mg treatment groups (-10.8%, P <.05 and -10.0%, P <.01, respectively). Uric acid levels were also significantly reduced at month 6 for both sibutramine treatment groups (10 mg, -6.0% [P <.05] and 15 mg, -6.2% [P <.05] compared with placebo, -1.9%). Decrease in uric acid was the only laboratory-reported variable that was statistically significantly greater in the sibutramine-treated patients than in the placebo-treated patients at end point Table 3. These changes are more closely associated with disease end points than with patient outcomes.

Among adverse events occurring with a frequency of more than 5% in any treatment group, only dry mouth occurred significantly more frequently with 10 mg sibutramine (19 patients) or 15 mg sibutramine (21 patients) than with placebo (2 patients; P <.001). Eight serious adverse events led to early withdrawal during the double-blind phase of the study (2 in each of the sibutramine groups and 4 in the placebo group). Two were considered possibly related to sibutramine. The first was a 49-year-old woman with a history of epilepsy with no seizures during the previous 4 years, who was withdrawn after 126 days of treatment with 10 mg sibutramine because she had had 4 “drop attacks” within 2 weeks. The second was a 32-year-old woman who was withdrawn after 246 days of treatment with 15 mg sibutramine because of palpitations due to frequent ventricular ectopic beats. No evidence was found of any withdrawal effect during the follow-up period, when active treatment was over. Patient psychological status was evaluated both at the end of the double-blind treatment (63 patients completed all assessments) and at follow-up 1 week after stopping treatment (67 patients completed all assessments). There were no statistically or clinically significant differences between the treatment groups for change in either the Beck Depression Inventory or the State Anxiety Inventory studied from the end of the double-blind treatment period to the follow-up. All 3 groups showed a mean improvement in the depression inventory.

Mean changes in vital signs averaged over all post-baseline double-blind assessments adjusted for center and treatment-by-center are shown in Table 3. Differences in mean changes in blood pressure were not statistically significant for the sibutramine treatment groups from placebo, except for the mean increase in diastolic blood pressure in the sibutramine 10 mg treatment group Table 3. Statistically significant changes in pulse rate in the sibutramine 15 mg treatment group compared with the placebo treatment group were consistent with the mechanism of action of sibutramine as an SNRI.

Discussion

Weight loss, obesity, and overweight are important topics,¹⁶ and today an increasing proportion of obesity is first encountered in the family practice setting.³⁸ Our study shows that in the context of a family practice setting, a patient’s diet, along with sibutramine treatment, will cause more loss of weight than changes in diet alone. Increased levels of total serum cholesterol are associated with an increased risk of coronary heart disease, and although difficult to quantify, a reduction of 0.6 mmol per L has been observed to be associated with a 54% lower risk.³⁹ This was calculated in a retrospective analysis of results of several clinical studies. Patients in this study losing 10% of their body weight over 6 months taking sibutramine 10 mg once daily reduced serum cholesterol by 0.23 mmol per L and may contribute to an overall benefit for patients as part of a risk reduction program. Similarly, the recently recommended target for reduction of triglycerides is a reduction of 1.7 mmol per L.⁴⁰ In our study, a reduction of 0.63 mmol per L in triglycerides was accomplished among those patients who received sibutramine 10 mg who lost at least 10% of body weight. Mean changes in blood pressure observed during the 6 months of our study were not great enough to increase the category of patient risk for coronary heart disease.⁴¹ Although weight loss can be associated with decreased mortality,¹³ the extent to which a weight reduction of 10% can augment other medical approaches to risk reduction long-term is only now being defined.

In this 1-year placebo-controlled study carried out in the primary care setting, sibutramine given with dietary advice was significantly superior in all weight loss efficacy assessments to dietary advice alone (placebo). These efficacy assessments included the ranked weight reduction outcomes analysis. Sibutramine was consistently superior to placebo across all categories of weight loss in this ranked analysis that takes subject withdrawals into account. The results of this study confirm the findings previously reported.³⁰

The plateau in weight loss observed at 6 months and maintained while having treatment is consistent with observations in other long-term studies of drug treatment in uncomplicated obese patients.^26,37 On the basis of other studies of weight loss, the regain of weight after cessation of treatment is not unexpected. The rates of completion of the study, although low, probably reflect what might be expected to occur in a general practice setting and are consistent with completion rates of weight loss programs in general.³⁷

The response rate at the 5% and 10% weight-reduction thresholds, with changes in waist circumference and in waist/hip ratio, were chosen for analysis in part because weight loss at these levels may ameliorate obesity-related disease.^37-42 Sustained moderate weight loss may be expected to bring about a reduction in blood pressure, hypertension, and dyslipidemia.⁴³ The small mean changes in blood pressure seen in the sibutramine groups were not unexpected given the norepinephrine-reuptake inhibitory properties of sibutramine. In clinical practice, any clinically significant changes that may occur should be detected by routine monitoring of vital signs. Of note, even in the current setting, where approximately 8% of patients were being treated with concomitant therapies related to cardiovascular disease, no patients were withdrawn (0%) because of elevated blood pressure.

Limitations

There were limitations to the general applicability of our study results to family practice. The treatment was limited to those patients who first demonstrated they could adhere to a weight loss program. Although the family practitioner may have some background information concerning this capability for a candidate for sibutramine treatment, general documentation may not be available for all candidates. Also, only 53% of the randomized patients completed this 1-year weight loss study. In addition, generally patients who stop weight loss treatment gain weight; however, the Sibutramine Trial in Obesity Reduction and Maintenance⁴⁴ demonstrated that weight loss can be maintained safely with sibutramine for up to 2 years. Another limitation to applicability of the study is the detailed depression and anxiety profile carried out for some patients enrolled. Such analysis is not practical or needed before sibutramine use in a family practice setting, and this analysis is not included in recently initiated studies. This was a generally healthy obese patient population with few comorbidities, and they were not depressed or anxious. That may limit the ability to generalize these findings to obese patients overall.

Conclusions

Dosages of sibutramine 10 mg or 15 mg administered once daily in a primary care general practice setting to patients with mild to moderate uncomplicated obesity produced statistically and clinically significant weight reduction that was maintained over a 12-month treatment period. Sibutramine was safe and well tolerated at both 10-mg and 15-mg once-daily doses. Further studies in patients with medical complications of obesity are under way to evaluate the efficacy of sibutramine-induced weight loss on the comorbidities and risk factors associated with obesity.

Acknowledgments

The author would like to thank the Sibutramine Clinical Study 1047 Team for their contributions, especially J Hosie, I James, and SP Jones. Competing interests: This study was funded by Knoll Pharmaceuticals, the manufacturer of sibutramine.

Obesity is a common problem, with significant health costs and associated morbidity; rates of overweight and obesity continue to increase in the United States^1,2 and Europe.³ In a typical general practice setting of 10,000 patients in the United Kingdom, approximately 1600 men and 1300 women could be expected to be overweight, and another 360 men and 700 women would need management of their obesity.⁴ Obesity predisposes patients to cardiovascular disease and hypertension,⁵ diabetes mellitus,⁶ hormone-related cancers,⁷ gallbladder disease,⁸ musculoskeletal disorders,⁹ sleep apnea,¹⁰ and generalized bodily pain.¹¹ It has a detrimental effect on social well-being,¹² decreases longevity,¹³ and imposes a tremendous financial and time burden on health care systems¹⁴; yet even a moderate amount of intentional weight loss (approximately 5%-10%) may significantly reduce the risk of mortality among subjects with medical complications of obesity.¹⁵

Increasingly, the major share of the primary effort to lower weight is falling on the general practice physician.¹⁶ Despite this, only 42% of obese patients seeking medical help are advised by their health professionals to lose weight; however, this telephone survey may suffer from a significant recall bias.¹⁷ Within the medical specialties, including family practice, internal medicine, and endocrinology, approximately 50% of physicians express interest in treating obesity in their practices.^16,18

Current treatment plans include mostly behavioral interventions, such as diet and exercise, along with behavior modification.¹⁹ These generally have a poor success rate, and in one study, long-term follow-up of weight maintenance after a 12-week very-low-calorie diet showed that patients regained the weight lost at a rate of 2.5% per month.²⁰ Although pharmacotherapy can be an important component of a weight loss program, safety issues have led to the withdrawal of some medications from the market. Studies with these have shown, however, that weight loss with pharmacotherapy can result in significant health benefits.^21-26

Sibutramine (Meridia; Knoll Pharmaceuticals, Nottingham, United Kingdom) is a new pharmacotherapy alternative. It is a serotonin-norepinephrine reuptake inhibitor (SNRI) that produces dose-related weight loss by enhancing postingestive satiety and by increasing energy expenditure.^21,27-29 Twelve months after weight reduction on a 4-week very-low-calorie diet, 75% of the patients taking sibutramine maintained 100% of their weight loss. This is a significantly greater percentage than the 42% who maintained 100% of weight loss in the nonsibutramine control group (P <.01).^30,31 Our study of sibutramine, carried out in a primary care general practice setting, was designed to determine the long-term efficacy, tolerability, and safety of dosages of 10 mg and 15 mg of sibutramine, which have been identified as effective and well tolerated in previous studies.^31,32 We also carried out analysis of improvements in risk factors resulting from 5% and 10% body weight loss.

Methods

The Nottinghamshire Independent Ethics Committee (Boots) approved our study, and all patients gave informed consent in writing. Our study was conducted in accordance with the Declaration of Helsinki (1983 revision).

Patients

We sought healthy male and female general practice patients, aged 18 to 65 years, who were mildly to moderately obese (body mass index [BMI] = 27 to 40 kg/m²), who had not lost more than 3 kg of weight during the previous 3 months, whose obesity was not of endocrine origin, and who did not have diabetes mellitus. All patients we enrolled were identified by primary care physicians. They had to have a seated pulse rate of 100 beats per minute or lower and a seated diastolic blood pressure of 100 mm Hg or lower. Hypertensive patients were included only if the condition had been stabilized with medication for 6 months. Patients receiving laxatives, anorectic agents, diuretics (except where stabilized for 6 months or more), bulking agents, antidepressants, or any other medication that in the opinion of the investigator might alter body weight were excluded. The patients completed the self-assessment Clinical Global Impressions questionnaire^32,33 and the Beck Depression Inventory. The objective of this was to identify patients who might be likely to have unusual weight changes in the context of the trial because of depression, and was not done for psychiatric evaluation. Those assessed by the investigator to be more than borderline depressed were excluded.

Trial Protocol

Because of the lack of a clear definition of obesity at the time of the initiation of our study and the lack of a clear definition of the patients for whom sibutramine treatment would be appropriate, protocol amendments permitted enrollment of a few patients with BMIs of 25 to 44 kg/m² (BMI Ž40 kg/m² is considered extremely obese). Depression and anxiety inventories were completed by a subset of patients at the study outset.³⁴ The investigator carried out a dietary inquiry for each eligible patient, and these patients were given standardized dietary advice and dietary advice sheets to be followed throughout the study. They were told to include certain foods in their diet each day: 12 oz of vegetables and fresh fruit; 6 oz of bread, cereal, potatoes, or rice; and 10 oz of skim milk; and they were told to substitute low-calorie foods such as fresh fruits and baked potatoes for sugary and fried foods such as chocolate and biscuits. The patients’ overall success in complying with dietary advice was assessed by the investigator using a 10-cm visual analog question scale labeled with opposite and extreme answers at either end (“easy to follow dietary advice” to “unable to follow dietary advice”).

Eligible patients then entered a 2-week single-blind placebo run-in period. At the end of the run-in period (month 0), those who still met the entry criteria entered the 12-month double-blind treatment phase of the study. Entry was restricted to those able to follow dietary advice as determined by the investigator. The patients were randomized in a double-blind fashion to once-daily treatment with placebo or sibutramine 10 mg or 15 mg dispensed in identical capsules, which were prepackaged and coded by the sponsor according to a computer-generated randomization list. The patients were assessed at monthly visits (month 0-month 12) during treatment and 1 week and 1 month after the completion of treatment. Patients withdrawing during treatment were assessed at that time and again 1 month after withdrawal. Beck Depression and State Anxiety inventories were performed on a subset of patients at the last treatment visit and at the 1-week follow-up visit to confirm lack of residual psychological changes due to medications.

At each visit, the investigator recorded the patient’s weight in indoor clothing. Waist and hip circumferences were measured at months 0, 6, and 12. Heart rate and blood pressure were recorded at every visit, and details of reported adverse events were recorded. Electrocardiograms and routine laboratory safety tests were performed at screening, 6 months, and 12 months.

Outcome Measures

The primary end point was the study outcome (measured on a categorical scale) on the basis of either the percentage weight loss at the end of the study or the reason for premature withdrawal.³⁵ The outcome measures were weight loss (kg), percentage weight loss, weight loss of at least 5% and at least 10%, change in BMI, and change in waist/hip ratio.

Statistical Analysis

Calculations were performed on the difference between treatment groups in weight change between baseline and month¹². Eighty-three patients completing treatment in each study group would be required to detect a difference in weight change of 4.0 kg with 90% power at the 5% significance level, assuming a standard deviation of 7.38 kg. Allowing for a forecast withdrawal rate of 45%, the desired sample size at enrollment was 184 patients per group or 552 for the entire study.

All statistical tests were two-tailed, with significance determined by reference to the 5% level. The null hypothesis was that the treatments were equivalent. The principal measure of efficacy was analyzed using the Kruskal-Wallis test,³⁶ including a factor for treatment group. Given that this result was significant at the 5% level, pairwise comparisons between the treatment groups were made by the large-sample normal approximation to the Wilcoxon rank sum test.³⁵

The methods we used to analyze differences between treatment groups were analysis of variance (ANOVA)³⁴ with factors for treatment group, center, and treatment group-by-center interaction for weight loss, waist and hip circumference, waist/hip ratio, vital signs, glucose and triglycerides, and ranked ANOVA for BMI, cholesterol, and uric acid. The differences between groups in the proportion of patients losing at least 5% and 10% of their baseline body weight were analyzed by logistic regression with factors for treatment group and center. All efficacy analyses were performed using the last observation carried forward (LOCF) to replace missing data. All data collected were included except for assessments performed more than 14 days after the last dose of study medication. After 14 days postdose, it was thought body weights could have changed significantly.

Results

Of 510 patients who entered the run-in period, 485 continued into the double-blind phase Figure 1. This occurred because of a lack of a gold-standard definition of obesity at the outset of this study; a protocol amendment allowed data from these patients to be included in the analysis. No clinically significant differences were noted between treatment groups at baseline with respect to demographic profile or characteristics of obesity Table 1. Four patients had BMIs at screening that fell out of the original trial protocol, all in the sibutramine 10-mg treatment group (1 patient <27 and 3 patients >40 kg/m²). Reanalysis of the data excluding these 4 patients was carried out, and no differences affecting the results, outcomes, or conclusions of our study were found.

Our analysis takes the patients who withdrew from the study into account.³² Pooled rates of withdrawal for any reason were similar in the 3 treatment groups: placebo, 83 patients (51%); sibutramine 10 mg, 79 (49%); and sibutramine 15 mg, 67 (42%). The completion rate for the 1-year study was 53%, which is generally consistent with weight loss programs.³⁷ Regarding the primary outcome analysis, fewer patients in the sibutramine groups than in the placebo group withdrew for lack of efficacy or adverse events. Using the log-rank test, there was no statistically significant difference between treatment groups in time to withdrawal, overall withdrawal rate, or withdrawal rate due to adverse events or lack of efficacy. Both sibutramine-treated groups had more patients in each category of weight loss above 5% (P=.001; Table 2. Pairwise comparisons showed a statistically significant difference versus placebo in favor of sibutramine 10 mg (P=.04) and sibutramine 15 mg (P <.001).

A significantly greater proportion of patients in the 2 sibutramine groups lost at least 5% and 10% of their baseline body weight Table 2. Mean weight reduction was significantly greater in the sibutramine groups than in the placebo group. Mean reductions in BMI reflected those in body weight and were significantly greater with 10 mg sibutramine (1.7 kg/m²) and 15 mg sibutramine (2.4 kg/m²) than with placebo (0.6 kg/m²; P <.001). The reduction in BMI was significantly greater with 15 mg sibutramine than 10-mg sibutramine (P <.05).

Patients treated with 10 mg or 15 mg sibutramine once daily lost more weight at each monthly assessment than those treated with placebo Figure 2. Weight loss was dose related, and maximal weight loss occurred by month 6 in all treatment groups. During the 4 weeks after treatment cessation (at whatever time this occurred), there were small weight increases in all treatment groups. Ten (6%) patients in the placebo group withdrew because of lack of efficacy compared with 5 (3%) in the sibutramine 10 mg group and 2 (1%) in the sibutramine 15-mg group.

Dose-related reductions in mean waist and hip circumferences achieved by month 6 were maintained at month 12. In the placebo group at month 12 (LOCF), the reductions were similar for waist and hip (2.4 cm and 2.6 cm, respectively), while in the 10- and 15-mg sibutramine groups, waist reduction was greater than hip reduction (6.4 cm and 3.8 cm at 10 mg and 7.4 cm and 5.2 cm at 15 mg). The reductions on active treatment were significantly greater than with placebo (P <.05 for overall comparison). The mean reduction in waist/hip ratio at month 12 (LOCF) was significantly greater with 10 mg sibutramine (-0.04) and 15 mg sibutramine (-0.03) than with placebo (-0.01; P <.05; least significant difference, .02).

There were statistically significant changes in triglycerides compared with placebo (+3.2%) at month 6 for the sibutramine 10- and 15-mg treatment groups (-10.8%, P <.05 and -10.0%, P <.01, respectively). Uric acid levels were also significantly reduced at month 6 for both sibutramine treatment groups (10 mg, -6.0% [P <.05] and 15 mg, -6.2% [P <.05] compared with placebo, -1.9%). Decrease in uric acid was the only laboratory-reported variable that was statistically significantly greater in the sibutramine-treated patients than in the placebo-treated patients at end point Table 3. These changes are more closely associated with disease end points than with patient outcomes.

Among adverse events occurring with a frequency of more than 5% in any treatment group, only dry mouth occurred significantly more frequently with 10 mg sibutramine (19 patients) or 15 mg sibutramine (21 patients) than with placebo (2 patients; P <.001). Eight serious adverse events led to early withdrawal during the double-blind phase of the study (2 in each of the sibutramine groups and 4 in the placebo group). Two were considered possibly related to sibutramine. The first was a 49-year-old woman with a history of epilepsy with no seizures during the previous 4 years, who was withdrawn after 126 days of treatment with 10 mg sibutramine because she had had 4 “drop attacks” within 2 weeks. The second was a 32-year-old woman who was withdrawn after 246 days of treatment with 15 mg sibutramine because of palpitations due to frequent ventricular ectopic beats. No evidence was found of any withdrawal effect during the follow-up period, when active treatment was over. Patient psychological status was evaluated both at the end of the double-blind treatment (63 patients completed all assessments) and at follow-up 1 week after stopping treatment (67 patients completed all assessments). There were no statistically or clinically significant differences between the treatment groups for change in either the Beck Depression Inventory or the State Anxiety Inventory studied from the end of the double-blind treatment period to the follow-up. All 3 groups showed a mean improvement in the depression inventory.

Mean changes in vital signs averaged over all post-baseline double-blind assessments adjusted for center and treatment-by-center are shown in Table 3. Differences in mean changes in blood pressure were not statistically significant for the sibutramine treatment groups from placebo, except for the mean increase in diastolic blood pressure in the sibutramine 10 mg treatment group Table 3. Statistically significant changes in pulse rate in the sibutramine 15 mg treatment group compared with the placebo treatment group were consistent with the mechanism of action of sibutramine as an SNRI.

Discussion

Weight loss, obesity, and overweight are important topics,¹⁶ and today an increasing proportion of obesity is first encountered in the family practice setting.³⁸ Our study shows that in the context of a family practice setting, a patient’s diet, along with sibutramine treatment, will cause more loss of weight than changes in diet alone. Increased levels of total serum cholesterol are associated with an increased risk of coronary heart disease, and although difficult to quantify, a reduction of 0.6 mmol per L has been observed to be associated with a 54% lower risk.³⁹ This was calculated in a retrospective analysis of results of several clinical studies. Patients in this study losing 10% of their body weight over 6 months taking sibutramine 10 mg once daily reduced serum cholesterol by 0.23 mmol per L and may contribute to an overall benefit for patients as part of a risk reduction program. Similarly, the recently recommended target for reduction of triglycerides is a reduction of 1.7 mmol per L.⁴⁰ In our study, a reduction of 0.63 mmol per L in triglycerides was accomplished among those patients who received sibutramine 10 mg who lost at least 10% of body weight. Mean changes in blood pressure observed during the 6 months of our study were not great enough to increase the category of patient risk for coronary heart disease.⁴¹ Although weight loss can be associated with decreased mortality,¹³ the extent to which a weight reduction of 10% can augment other medical approaches to risk reduction long-term is only now being defined.

In this 1-year placebo-controlled study carried out in the primary care setting, sibutramine given with dietary advice was significantly superior in all weight loss efficacy assessments to dietary advice alone (placebo). These efficacy assessments included the ranked weight reduction outcomes analysis. Sibutramine was consistently superior to placebo across all categories of weight loss in this ranked analysis that takes subject withdrawals into account. The results of this study confirm the findings previously reported.³⁰

The plateau in weight loss observed at 6 months and maintained while having treatment is consistent with observations in other long-term studies of drug treatment in uncomplicated obese patients.^26,37 On the basis of other studies of weight loss, the regain of weight after cessation of treatment is not unexpected. The rates of completion of the study, although low, probably reflect what might be expected to occur in a general practice setting and are consistent with completion rates of weight loss programs in general.³⁷

The response rate at the 5% and 10% weight-reduction thresholds, with changes in waist circumference and in waist/hip ratio, were chosen for analysis in part because weight loss at these levels may ameliorate obesity-related disease.^37-42 Sustained moderate weight loss may be expected to bring about a reduction in blood pressure, hypertension, and dyslipidemia.⁴³ The small mean changes in blood pressure seen in the sibutramine groups were not unexpected given the norepinephrine-reuptake inhibitory properties of sibutramine. In clinical practice, any clinically significant changes that may occur should be detected by routine monitoring of vital signs. Of note, even in the current setting, where approximately 8% of patients were being treated with concomitant therapies related to cardiovascular disease, no patients were withdrawn (0%) because of elevated blood pressure.

Limitations

There were limitations to the general applicability of our study results to family practice. The treatment was limited to those patients who first demonstrated they could adhere to a weight loss program. Although the family practitioner may have some background information concerning this capability for a candidate for sibutramine treatment, general documentation may not be available for all candidates. Also, only 53% of the randomized patients completed this 1-year weight loss study. In addition, generally patients who stop weight loss treatment gain weight; however, the Sibutramine Trial in Obesity Reduction and Maintenance⁴⁴ demonstrated that weight loss can be maintained safely with sibutramine for up to 2 years. Another limitation to applicability of the study is the detailed depression and anxiety profile carried out for some patients enrolled. Such analysis is not practical or needed before sibutramine use in a family practice setting, and this analysis is not included in recently initiated studies. This was a generally healthy obese patient population with few comorbidities, and they were not depressed or anxious. That may limit the ability to generalize these findings to obese patients overall.

Conclusions

Dosages of sibutramine 10 mg or 15 mg administered once daily in a primary care general practice setting to patients with mild to moderate uncomplicated obesity produced statistically and clinically significant weight reduction that was maintained over a 12-month treatment period. Sibutramine was safe and well tolerated at both 10-mg and 15-mg once-daily doses. Further studies in patients with medical complications of obesity are under way to evaluate the efficacy of sibutramine-induced weight loss on the comorbidities and risk factors associated with obesity.

Acknowledgments

The author would like to thank the Sibutramine Clinical Study 1047 Team for their contributions, especially J Hosie, I James, and SP Jones. Competing interests: This study was funded by Knoll Pharmaceuticals, the manufacturer of sibutramine.